03_pytorch_computer_vision.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "c2dc16c4-c7f4-4945-ba91-6430a51e6f5a",
   "metadata": {
    "id": "c2dc16c4-c7f4-4945-ba91-6430a51e6f5a"
   },
   "source": [
    "<a href=\"https://colab.research.google.com/github/mrdbourke/pytorch-deep-learning/blob/main/03_pytorch_computer_vision.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>\n",
    "\n",
    "[View Source Code](https://github.com/mrdbourke/pytorch-deep-learning/blob/main/03_pytorch_computer_vision.ipynb) | [View Slides](https://github.com/mrdbourke/pytorch-deep-learning/blob/main/slides/03_pytorch_computer_vision.pdf) | [Watch Video Walkthrough](https://youtu.be/Z_ikDlimN6A?t=50417) "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "08f47c6a-3318-4e3f-8bb3-c520e00e63dd",
   "metadata": {
    "id": "08f47c6a-3318-4e3f-8bb3-c520e00e63dd"
   },
   "source": [
    "# 03. PyTorch Computer Vision\n",
    "\n",
    "[Computer vision](https://en.wikipedia.org/wiki/Computer_vision) is the art of teaching a computer to see.\n",
    "\n",
    "For example, it could involve building a model to classify whether a photo is of a cat or a dog ([binary classification](https://developers.google.com/machine-learning/glossary#binary-classification)).\n",
    "\n",
    "Or whether a photo is of a cat, dog or chicken ([multi-class classification](https://developers.google.com/machine-learning/glossary#multi-class-classification)).\n",
    "\n",
    "Or identifying where a car appears in a video frame ([object detection](https://en.wikipedia.org/wiki/Object_detection)).\n",
    "\n",
    "Or figuring out where different objects in an image can be separated ([panoptic segmentation](https://arxiv.org/abs/1801.00868)).\n",
    "\n",
    "![example computer vision problems](https://raw.githubusercontent.com/mrdbourke/pytorch-deep-learning/main/images/03-computer-vision-problems.png)\n",
    "*Example computer vision problems for binary classification, multiclass classification, object detection and segmentation.*"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "19179a39-0c6c-40f7-9891-09e17d107ecf",
   "metadata": {
    "id": "19179a39-0c6c-40f7-9891-09e17d107ecf"
   },
   "source": [
    "## Where does computer vision get used?\n",
    "\n",
    "If you use a smartphone, you've already used computer vision.\n",
    "\n",
    "Camera and photo apps use [computer vision to enhance](https://machinelearning.apple.com/research/panoptic-segmentation) and sort images.\n",
    "\n",
    "Modern cars use [computer vision](https://youtu.be/j0z4FweCy4M?t=2989) to avoid other cars and stay within lane lines.\n",
    "\n",
    "Manufacturers use computer vision to identify defects in various products.\n",
    "\n",
    "Security cameras use computer vision to detect potential intruders.\n",
    "\n",
    "In essence, anything that can described in a visual sense can be a potential computer vision problem."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "412e8bd1-0e6b-4ad6-8506-b28a8f669dc1",
   "metadata": {
    "id": "412e8bd1-0e6b-4ad6-8506-b28a8f669dc1"
   },
   "source": [
    "## What we're going to cover\n",
    "\n",
    "We're going to apply the PyTorch Workflow we've been learning in the past couple of sections to computer vision.\n",
    "\n",
    "![a PyTorch workflow with a computer vision focus](https://raw.githubusercontent.com/mrdbourke/pytorch-deep-learning/main/images/03-pytorch-computer-vision-workflow.png)\n",
    "\n",
    "Specifically, we're going to cover:\n",
    "\n",
    "| **Topic** | **Contents** |\n",
    "| ----- | ----- |\n",
    "| **0. Computer vision libraries in PyTorch** | PyTorch has a bunch of built-in helpful computer vision libraries, let's check them out.  |\n",
    "| **1. Load data** | To practice computer vision, we'll start with some images of different pieces of clothing from [FashionMNIST](https://github.com/zalandoresearch/fashion-mnist). |\n",
    "| **2. Prepare data** | We've got some images, let's load them in with a [PyTorch `DataLoader`](https://pytorch.org/docs/stable/data.html) so we can use them with our training loop. |\n",
    "| **3. Model 0: Building a baseline model** | Here we'll create a multi-class classification model to learn patterns in the data, we'll also choose a **loss function**, **optimizer** and build a **training loop**. | \n",
    "| **4. Making predictions and evaluting model 0** | Let's make some predictions with our baseline model and evaluate them. |\n",
    "| **5. Setup device agnostic code for future models** | It's best practice to write device-agnostic code, so let's set it up. |\n",
    "| **6. Model 1: Adding non-linearity** | Experimenting is a large part of machine learning, let's try and improve upon our baseline model by adding non-linear layers. |\n",
    "| **7. Model 2: Convolutional Neural Network (CNN)** | Time to get computer vision specific and introduce the powerful convolutional neural network architecture. |\n",
    "| **8. Comparing our models** | We've built three different models, let's compare them. |\n",
    "| **9. Evaluating our best model** | Let's make some predictons on random images and evaluate our best model. |\n",
    "| **10. Making a confusion matrix** | A confusion matrix is a great way to evaluate a classification model, let's see how we can make one. |\n",
    "| **11. Saving and loading the best performing model** | Since we might want to use our model for later, let's save it and make sure it loads back in correctly. |"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "cddf62c3-f5e5-4f7e-852a-2ad6d38b7399",
   "metadata": {
    "id": "cddf62c3-f5e5-4f7e-852a-2ad6d38b7399"
   },
   "source": [
    "## Where can can you get help?\n",
    "\n",
    "All of the materials for this course [live on GitHub](https://github.com/mrdbourke/pytorch-deep-learning).\n",
    "\n",
    "If you run into trouble, you can ask a question on the course [GitHub Discussions page](https://github.com/mrdbourke/pytorch-deep-learning/discussions) there too.\n",
    "\n",
    "And of course, there's the [PyTorch documentation](https://pytorch.org/docs/stable/index.html) and [PyTorch developer forums](https://discuss.pytorch.org/), a very helpful place for all things PyTorch. "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a0bedcfc-e12a-4a81-9913-84c6a888742a",
   "metadata": {
    "id": "a0bedcfc-e12a-4a81-9913-84c6a888742a"
   },
   "source": [
    "## 0. Computer vision libraries in PyTorch\n",
    "\n",
    "Before we get started writing code, let's talk about some PyTorch computer vision libraries you should be aware of.\n",
    "\n",
    "| PyTorch module | What does it do? |\n",
    "| ----- | ----- |\n",
    "| [`torchvision`](https://pytorch.org/vision/stable/index.html) | Contains datasets, model architectures and image transformations often used for computer vision problems. |\n",
    "| [`torchvision.datasets`](https://pytorch.org/vision/stable/datasets.html) | Here you'll find many example computer vision datasets for a range of problems from image classification, object detection, image captioning, video classification and more. It also contains [a series of base classes for making custom datasets](https://pytorch.org/vision/stable/datasets.html#base-classes-for-custom-datasets). |\n",
    "| [`torchvision.models`](https://pytorch.org/vision/stable/models.html) | This module contains well-performing and commonly used computer vision model architectures implemented in PyTorch, you can use these with your own problems. | \n",
    "| [`torchvision.transforms`](https://pytorch.org/vision/stable/transforms.html) | Often images need to be transformed (turned into numbers/processed/augmented) before being used with a model, common image transformations are found here. | \n",
    "| [`torch.utils.data.Dataset`](https://pytorch.org/docs/stable/data.html#torch.utils.data.Dataset) | Base dataset class for PyTorch.  | \n",
    "| [`torch.utils.data.DataLoader`](https://pytorch.org/docs/stable/data.html#module-torch.utils.data) | Creates a Python iterable over a dataset (created with `torch.utils.data.Dataset`). |\n",
    "\n",
    "> **Note:** The `torch.utils.data.Dataset` and `torch.utils.data.DataLoader` classes aren't only for computer vision in PyTorch, they are capable of dealing with many different types of data.\n",
    "\n",
    "Now we've covered some of the most important PyTorch computer vision libraries, let's import the relevant dependencies.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "c263a60d-d788-482f-b9e7-9cab4f6b1f72",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "c263a60d-d788-482f-b9e7-9cab4f6b1f72",
    "outputId": "20ba933b-6026-475f-a8d9-12cf416aff74"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "PyTorch version: 2.4.1\n",
      "torchvision version: 0.19.1\n"
     ]
    }
   ],
   "source": [
    "# Import PyTorch\n",
    "import torch\n",
    "from torch import nn\n",
    "\n",
    "# Import torchvision \n",
    "import torchvision\n",
    "from torchvision import datasets\n",
    "from torchvision.transforms import ToTensor\n",
    "\n",
    "# Import matplotlib for visualization\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "# Check versions\n",
    "# Note: your PyTorch version shouldn't be lower than 1.10.0 and torchvision version shouldn't be lower than 0.11\n",
    "print(f\"PyTorch version: {torch.__version__}\\ntorchvision version: {torchvision.__version__}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "48d6bfe7-91da-44eb-9ab6-7c41c1e9fa8e",
   "metadata": {
    "id": "48d6bfe7-91da-44eb-9ab6-7c41c1e9fa8e"
   },
   "source": [
    "## 1. Getting a dataset\n",
    "\n",
    "To begin working on a computer vision problem, let's get a computer vision dataset.\n",
    "\n",
    "We're going to start with FashionMNIST.\n",
    "\n",
    "MNIST stands for Modified National Institute of Standards and Technology.\n",
    "\n",
    "The [original MNIST dataset](https://en.wikipedia.org/wiki/MNIST_database) contains thousands of examples of handwritten digits (from 0 to 9) and was used to build computer vision models to identify numbers for postal services.\n",
    "\n",
    "[FashionMNIST](https://github.com/zalandoresearch/fashion-mnist), made by Zalando Research, is a similar setup. \n",
    "\n",
    "Except it contains grayscale images of 10 different kinds of clothing.\n",
    "\n",
    "![example image of FashionMNIST](https://raw.githubusercontent.com/mrdbourke/pytorch-deep-learning/main/images/03-fashion-mnist-slide.png)\n",
    "*`torchvision.datasets` contains a lot of example datasets you can use to practice writing computer vision code on. FashionMNIST is one of those datasets. And since it has 10 different image classes (different types of clothing), it's a multi-class classification problem.*\n",
    "\n",
    "Later, we'll be building a computer vision neural network to identify the different styles of clothing in these images.\n",
    "\n",
    "PyTorch has a bunch of common computer vision datasets stored in `torchvision.datasets`.\n",
    "\n",
    "Including FashionMNIST in [`torchvision.datasets.FashionMNIST()`](https://pytorch.org/vision/main/generated/torchvision.datasets.FashionMNIST.html).\n",
    "\n",
    "To download it, we provide the following parameters:\n",
    "* `root: str` - which folder do you want to download the data to?\n",
    "* `train: Bool` - do you want the training or test split?\n",
    "* `download: Bool` - should the data be downloaded?\n",
    "* `transform: torchvision.transforms` - what transformations would you like to do on the data?\n",
    "* `target_transform` - you can transform the targets (labels) if you like too.\n",
    "\n",
    "Many other datasets in `torchvision` have these parameter options."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "7a646459",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Operations to get the mean and std of the training dataset\n",
    "import torchvision.transforms.v2 as transforms\n",
    "from torch.utils.data import DataLoader\n",
    "\n",
    "# Optional, just to normalize with respect to the mean and std\n",
    "\n",
    "trans = transforms.Compose([transforms.ToImage(),\n",
    "                            transforms.ToDtype(torch.float32, scale=True)],\n",
    "                             )\n",
    "\n",
    "train_data_t = datasets.FashionMNIST(\n",
    "    root=\"data\", # where to download data to?\n",
    "    train=True, # get training data\n",
    "    download=True, # download data if it doesn't exist on disk\n",
    "    transform=trans, # images come as PIL format, we want to turn into Torch tensors\n",
    "    target_transform=None # you can transform labels as well\n",
    ")\n",
    "stats_loader = DataLoader(train_data_t, batch_size=len(train_data_t), shuffle=True)\n",
    "images, _ = next(iter(stats_loader))\n",
    "data_mean, data_std = torch.mean(images), torch.std(images)\n",
    "\n",
    "trans = transforms.Compose([transforms.ToImage(),\n",
    "                            transforms.ToDtype(torch.float32, scale=True),\n",
    "                            transforms.Normalize((data_mean,), (data_std,))]\n",
    "                            )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "486f8377-6810-4367-859d-69dccc7aef95",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "486f8377-6810-4367-859d-69dccc7aef95",
    "outputId": "877f93b2-12c5-477e-92bf-3ec3f1449282"
   },
   "outputs": [],
   "source": [
    "# Setup training data\n",
    "train_data = datasets.FashionMNIST(\n",
    "    root=\"data\", # where to download data to?\n",
    "    train=True, # get training data\n",
    "    download=True, # download data if it doesn't exist on disk\n",
    "    transform=trans, #ToTensor() # images come as PIL format, we want to turn into Torch tensors\n",
    "    target_transform=None # you can transform labels as well\n",
    ")\n",
    "\n",
    "# Setup testing data\n",
    "test_data = datasets.FashionMNIST(\n",
    "    root=\"data\",\n",
    "    train=False, # get test data\n",
    "    download=True,\n",
    "    transform=trans # ToTensor()\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a63246f6-3645-49de-88fe-ec18e78bfbaf",
   "metadata": {
    "id": "a63246f6-3645-49de-88fe-ec18e78bfbaf"
   },
   "source": [
    "Let's check out the first sample of the training data."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "43bfd3d9-a132-41e8-8ccd-5ae25a7da59a",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "43bfd3d9-a132-41e8-8ccd-5ae25a7da59a",
    "outputId": "1595e80b-6a3f-4171-a128-ec506b4d8326"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(Image([[[-8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01,\n",
       "          -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01,\n",
       "          -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01,\n",
       "          -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01,\n",
       "          -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01,\n",
       "          -8.1026e-01, -8.1026e-01, -8.1026e-01],\n",
       "         [-8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01,\n",
       "          -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01,\n",
       "          -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01,\n",
       "          -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01,\n",
       "          -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01,\n",
       "          -8.1026e-01, -8.1026e-01, -8.1026e-01],\n",
       "         [-8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01,\n",
       "          -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01,\n",
       "          -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01,\n",
       "          -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01,\n",
       "          -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01,\n",
       "          -8.1026e-01, -8.1026e-01, -8.1026e-01],\n",
       "         [-8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01,\n",
       "          -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01,\n",
       "          -8.1026e-01, -8.1026e-01, -7.9915e-01, -8.1026e-01, -8.1026e-01,\n",
       "          -6.6585e-01,  6.6261e-04, -8.1026e-01, -8.1026e-01, -7.9915e-01,\n",
       "          -7.6582e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01,\n",
       "          -7.9915e-01, -7.9915e-01, -8.1026e-01],\n",
       "         [-8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01,\n",
       "          -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01,\n",
       "          -8.1026e-01, -8.1026e-01, -7.7693e-01, -8.1026e-01, -4.1035e-01,\n",
       "           7.0050e-01,  6.0052e-01, -1.2153e-01, -2.1040e-01, -8.1026e-01,\n",
       "          -8.1026e-01, -8.1026e-01, -7.9915e-01, -7.7693e-01, -7.6582e-01,\n",
       "          -8.1026e-01, -8.1026e-01, -7.7693e-01],\n",
       "         [-8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01,\n",
       "          -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01,\n",
       "          -8.1026e-01, -8.1026e-01, -7.4361e-01, -8.1026e-01,  3.2281e-01,\n",
       "           1.4559e+00,  1.1448e+00,  6.7828e-01,  7.8937e-01,  5.5609e-01,\n",
       "          -5.5476e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01,\n",
       "          -6.7696e-01, -6.9917e-01, -8.1026e-01],\n",
       "         [-8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01,\n",
       "          -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01,\n",
       "          -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01,  9.1156e-01,\n",
       "           1.8113e+00,  1.4892e+00,  1.1671e+00,  3.7835e-01,  9.2267e-01,\n",
       "           9.7821e-01,  4.0057e-01, -9.9314e-02, -5.5476e-01,  4.5097e-02,\n",
       "           6.3385e-01, -1.0446e-02, -6.4363e-01],\n",
       "         [-8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01,\n",
       "          -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01,\n",
       "          -8.1026e-01, -7.9915e-01, -8.1026e-01, -4.3771e-02,  1.4892e+00,\n",
       "           1.6669e+00,  1.6114e+00,  1.5892e+00,  1.5892e+00,  1.0004e+00,\n",
       "           6.0052e-01,  5.3387e-01,  5.4498e-01,  8.1158e-01,  7.5604e-01,\n",
       "           1.6729e-01,  1.1004e+00, -7.7097e-02],\n",
       "         [-8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01,\n",
       "          -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01, -7.9915e-01,\n",
       "          -7.9915e-01, -7.9915e-01, -8.1026e-01,  1.4114e+00,  1.7669e+00,\n",
       "           1.7669e+00,  1.7780e+00,  1.7336e+00,  1.6669e+00,  1.6669e+00,\n",
       "           1.5781e+00,  1.5559e+00,  1.0115e+00,  6.0052e-01,  5.5609e-01,\n",
       "           1.3670e+00,  1.7336e+00, -8.1026e-01],\n",
       "         [-8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01,\n",
       "          -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01,\n",
       "          -8.1026e-01, -8.1026e-01, -8.1026e-01,  1.2226e+00,  1.6892e+00,\n",
       "           1.5892e+00,  1.6669e+00,  1.7225e+00,  1.8002e+00,  1.7114e+00,\n",
       "           1.6780e+00,  1.6558e+00,  1.6780e+00,  1.6447e+00,  1.6669e+00,\n",
       "           1.9113e+00,  1.1115e+00, -8.1026e-01],\n",
       "         [-8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01,\n",
       "          -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01,\n",
       "          -8.1026e-01, -8.1026e-01, -8.1026e-01,  1.3337e+00,  1.7225e+00,\n",
       "           1.6114e+00,  1.5559e+00,  1.3892e+00,  1.1893e+00,  1.5447e+00,\n",
       "           1.5225e+00,  1.5336e+00,  1.5559e+00,  1.6669e+00,  1.6336e+00,\n",
       "           1.8891e+00,  1.4337e+00, -8.1026e-01],\n",
       "         [-8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01,\n",
       "          -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01, -7.9915e-01,\n",
       "          -7.7693e-01, -8.1026e-01, -6.7696e-01,  1.6225e+00,  1.6336e+00,\n",
       "           1.5447e+00,  1.6114e+00,  1.3226e+00,  1.0671e+00,  1.7114e+00,\n",
       "           1.5003e+00,  1.6114e+00,  1.6780e+00,  1.5447e+00,  1.7003e+00,\n",
       "           1.3781e+00,  1.5114e+00, -2.3262e-01],\n",
       "         [-8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01,\n",
       "          -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01,\n",
       "          -7.4361e-01, -8.1026e-01,  2.8948e-01,  1.9002e+00,  1.6558e+00,\n",
       "           1.6336e+00,  1.6114e+00,  1.4448e+00,  1.3892e+00,  1.6447e+00,\n",
       "           1.5781e+00,  1.5559e+00,  1.6558e+00,  1.6336e+00,  1.9113e+00,\n",
       "           5.1165e-01,  1.0449e+00, -1.8818e-01],\n",
       "         [-8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01,\n",
       "          -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01, -7.6582e-01,\n",
       "          -8.1026e-01, -8.1026e-01, -1.9929e-01,  1.8113e+00,  1.7225e+00,\n",
       "           1.7447e+00,  1.7225e+00,  1.8558e+00,  1.7669e+00,  1.5559e+00,\n",
       "           1.6114e+00,  1.6669e+00,  1.7891e+00,  1.6003e+00,  1.6003e+00,\n",
       "           1.5114e+00,  2.1172e-01, -8.1026e-01],\n",
       "         [-8.1026e-01, -8.1026e-01, -7.9915e-01, -7.6582e-01, -7.4361e-01,\n",
       "          -7.3250e-01, -7.8804e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01,\n",
       "          -8.1026e-01, -8.1026e-01,  1.8225e+00,  1.7003e+00,  1.6003e+00,\n",
       "           1.6669e+00,  1.6558e+00,  1.6225e+00,  1.6558e+00,  1.6447e+00,\n",
       "           1.5892e+00,  1.6669e+00,  1.7336e+00,  1.5781e+00,  1.6114e+00,\n",
       "           2.0224e+00,  4.5097e-02, -8.1026e-01],\n",
       "         [-8.1026e-01, -7.7693e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01,\n",
       "          -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01, -1.2153e-01,\n",
       "           8.0047e-01,  1.4559e+00,  1.7225e+00,  1.4892e+00,  1.5559e+00,\n",
       "           1.6447e+00,  1.6114e+00,  1.5003e+00,  1.5336e+00,  1.6114e+00,\n",
       "           1.6780e+00,  1.6669e+00,  1.6225e+00,  1.5781e+00,  1.6780e+00,\n",
       "           1.9002e+00,  9.5599e-01, -8.1026e-01],\n",
       "         [-8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01, -6.1030e-01,\n",
       "          -3.2148e-01,  1.0064e-01,  3.7835e-01,  1.2892e+00,  1.7225e+00,\n",
       "           1.6336e+00,  1.6558e+00,  1.6003e+00,  1.7003e+00,  1.4114e+00,\n",
       "           1.4670e+00,  1.5336e+00,  1.7447e+00,  1.6780e+00,  1.7891e+00,\n",
       "           1.1448e+00,  1.2781e+00,  1.9669e+00,  1.9446e+00,  1.7780e+00,\n",
       "           1.8336e+00,  1.5781e+00, -8.1026e-01],\n",
       "         [-8.1026e-01, -1.7707e-01,  1.2670e+00,  1.5003e+00,  1.6780e+00,\n",
       "           1.6447e+00,  1.6780e+00,  1.5003e+00,  1.4559e+00,  1.5670e+00,\n",
       "           1.5003e+00,  1.5114e+00,  1.4114e+00,  9.5599e-01,  1.9113e+00,\n",
       "           1.3337e+00,  1.4781e+00,  1.6669e+00,  2.0224e+00,  2.0224e+00,\n",
       "           1.6447e+00,  1.7891e+00,  1.6447e+00,  1.5336e+00,  1.6336e+00,\n",
       "           1.7669e+00,  1.9224e+00, -8.1026e-01],\n",
       "         [-7.7693e-01,  1.4337e+00,  1.7225e+00,  1.6780e+00,  1.6447e+00,\n",
       "           1.5336e+00,  1.5336e+00,  1.5670e+00,  1.4670e+00,  1.4670e+00,\n",
       "           1.4670e+00,  1.6336e+00,  1.8558e+00,  7.8422e-02,  8.5602e-01,\n",
       "           2.0224e+00,  1.7336e+00,  1.6447e+00,  1.2781e+00,  9.0045e-01,\n",
       "           1.3115e+00,  1.5225e+00,  1.4559e+00,  1.5114e+00,  1.6558e+00,\n",
       "           1.7225e+00,  1.6892e+00, -8.1026e-01],\n",
       "         [ 2.7838e-01,  1.7780e+00,  1.3892e+00,  1.5225e+00,  1.6558e+00,\n",
       "           1.7336e+00,  1.7336e+00,  1.7891e+00,  1.9558e+00,  1.6336e+00,\n",
       "           1.3448e+00,  1.5781e+00,  1.6003e+00,  1.8669e+00, -8.8205e-02,\n",
       "           6.6261e-04,  3.6724e-01,  4.8944e-01,  1.0560e+00,  1.6225e+00,\n",
       "           1.6447e+00,  1.5781e+00,  1.6003e+00,  1.6669e+00,  1.6669e+00,\n",
       "           1.6780e+00,  1.7336e+00, -4.8811e-01],\n",
       "         [ 2.2880e-02,  1.4559e+00,  1.5447e+00,  1.4559e+00,  1.3337e+00,\n",
       "           1.4670e+00,  1.5336e+00,  1.6892e+00,  1.5892e+00,  1.2448e+00,\n",
       "           1.3781e+00,  1.4781e+00,  1.3892e+00,  1.5559e+00,  1.8558e+00,\n",
       "           1.3559e+00,  1.7114e+00,  1.9113e+00,  1.8447e+00,  1.6669e+00,\n",
       "           1.6114e+00,  1.5447e+00,  1.5114e+00,  1.6558e+00,  1.6336e+00,\n",
       "           1.6447e+00,  1.7447e+00, -6.5988e-02],\n",
       "         [-2.7705e-01,  1.4448e+00,  1.2226e+00,  1.3448e+00,  1.5559e+00,\n",
       "           1.3781e+00,  1.2448e+00,  1.3004e+00,  1.3448e+00,  1.3226e+00,\n",
       "           1.4337e+00,  1.5670e+00,  1.6225e+00,  1.6447e+00,  1.6336e+00,\n",
       "           1.8113e+00,  1.6892e+00,  1.5892e+00,  1.4003e+00,  1.4781e+00,\n",
       "           1.2559e+00,  1.2004e+00,  1.1559e+00,  1.1004e+00,  1.2004e+00,\n",
       "           1.4670e+00,  1.4781e+00,  4.6722e-01],\n",
       "         [-8.1026e-01,  5.4498e-01,  1.6225e+00,  1.3337e+00,  1.1782e+00,\n",
       "           1.0893e+00,  1.2226e+00,  1.3670e+00,  1.4559e+00,  1.5225e+00,\n",
       "           1.5559e+00,  1.4892e+00,  1.5336e+00,  1.5225e+00,  1.4114e+00,\n",
       "           1.3670e+00,  1.3448e+00,  1.3115e+00,  1.3559e+00,  1.3115e+00,\n",
       "           1.3892e+00,  1.3226e+00,  1.1448e+00,  9.2267e-01,  1.0449e+00,\n",
       "           1.1559e+00,  1.5225e+00,  2.1172e-01],\n",
       "         [-8.1026e-01, -8.1026e-01,  1.1771e-02,  1.2892e+00,  1.5447e+00,\n",
       "           1.3115e+00,  1.1337e+00,  1.1004e+00,  1.1337e+00,  1.2004e+00,\n",
       "           1.2448e+00,  1.2781e+00,  1.2892e+00,  1.2781e+00,  1.3337e+00,\n",
       "           1.3892e+00,  1.4559e+00,  1.5114e+00,  1.5225e+00,  1.5225e+00,\n",
       "           1.5336e+00,  1.2781e+00,  1.2781e+00,  1.3448e+00,  1.3226e+00,\n",
       "           1.5892e+00,  1.0782e+00, -8.1026e-01],\n",
       "         [-7.8804e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01, -7.7097e-02,\n",
       "           1.4114e+00,  1.6558e+00,  1.8225e+00,  1.8447e+00,  1.8780e+00,\n",
       "           1.9224e+00,  1.8891e+00,  1.9002e+00,  1.6447e+00,  1.6336e+00,\n",
       "           1.3337e+00,  1.3115e+00,  1.1782e+00,  1.2115e+00,  1.2115e+00,\n",
       "           1.2004e+00,  1.1448e+00,  1.0338e+00,  1.0560e+00,  2.8948e-01,\n",
       "          -1.6596e-01, -8.1026e-01, -8.1026e-01],\n",
       "         [-8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01,\n",
       "          -8.1026e-01, -8.1026e-01, -3.6592e-01, -1.3264e-01, -3.2148e-01,\n",
       "          -1.0446e-02, -3.5481e-01, -4.2146e-01, -8.1026e-01, -8.1026e-01,\n",
       "          -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01,\n",
       "          -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01,\n",
       "          -8.1026e-01, -8.1026e-01, -8.1026e-01],\n",
       "         [-8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01,\n",
       "          -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01,\n",
       "          -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01,\n",
       "          -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01,\n",
       "          -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01,\n",
       "          -8.1026e-01, -8.1026e-01, -8.1026e-01],\n",
       "         [-8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01,\n",
       "          -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01,\n",
       "          -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01,\n",
       "          -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01,\n",
       "          -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01, -8.1026e-01,\n",
       "          -8.1026e-01, -8.1026e-01, -8.1026e-01]]], ),\n",
       " 9,\n",
       " 60000)"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# See first training sample\n",
    "image, label = train_data[0]\n",
    "image, label, len(train_data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "8400d003",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'T-shirt/top': 0,\n",
       " 'Trouser': 1,\n",
       " 'Pullover': 2,\n",
       " 'Dress': 3,\n",
       " 'Coat': 4,\n",
       " 'Sandal': 5,\n",
       " 'Shirt': 6,\n",
       " 'Sneaker': 7,\n",
       " 'Bag': 8,\n",
       " 'Ankle boot': 9}"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "class_to_idx = train_data.class_to_idx\n",
    "class_to_idx"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9ad9d782-06cb-4591-ae3c-3a8b2389a1b2",
   "metadata": {
    "id": "9ad9d782-06cb-4591-ae3c-3a8b2389a1b2"
   },
   "source": [
    "### 1.1 Input and output shapes of a computer vision model\n",
    "\n",
    "We've got a big tensor of values (the image) leading to a single value for the target (the label).\n",
    "\n",
    "Let's see the image shape."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "c2997d9f-b574-4d23-aa34-1a4df1751226",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "c2997d9f-b574-4d23-aa34-1a4df1751226",
    "outputId": "d9c4283b-aab8-410f-dd7f-03f08d943366"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "torch.Size([1, 28, 28])"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# What's the shape of the image?\n",
    "image.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "c6a0a8bb",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "9"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "label"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b5326a05-f807-448d-99a3-6d03fc8739f8",
   "metadata": {
    "id": "b5326a05-f807-448d-99a3-6d03fc8739f8"
   },
   "source": [
    "The shape of the image tensor is `[1, 28, 28]` or more specifically:\n",
    "\n",
    "```\n",
    "[color_channels=1, height=28, width=28]\n",
    "```\n",
    "\n",
    "Having `color_channels=1` means the image is grayscale.\n",
    "\n",
    "![example input and output shapes of the fashionMNIST problem](https://raw.githubusercontent.com/mrdbourke/pytorch-deep-learning/main/images/03-computer-vision-input-and-output-shapes.png)\n",
    "*Various problems will have various input and output shapes. But the premise remains: encode data into numbers, build a model to find patterns in those numbers, convert those patterns into something meaningful.*\n",
    "\n",
    "If `color_channels=3`, the image comes in pixel values for red, green and blue (this is also known a the [RGB color model](https://en.wikipedia.org/wiki/RGB_color_model)).\n",
    "\n",
    "The order of our current tensor is often referred to as `CHW` (Color Channels, Height, Width).\n",
    "\n",
    "There's debate on whether images should be represented as `CHW` (color channels first) or `HWC` (color channels last).\n",
    "\n",
    "> **Note:** You'll also see `NCHW` and `NHWC` formats where `N` stands for *number of images*. For example if you have a `batch_size=32`, your tensor shape may be `[32, 1, 28, 28]`. We'll cover batch sizes later.\n",
    "\n",
    "PyTorch generally accepts `NCHW` (channels first) as the default for many operators.\n",
    "\n",
    "However, PyTorch also explains that `NHWC` (channels last) performs better and is [considered best practice](https://pytorch.org/blog/tensor-memory-format-matters/#pytorch-best-practice). \n",
    "\n",
    "For now, since our dataset and models are relatively small, this won't make too much of a difference.\n",
    "\n",
    "But keep it in mind for when you're working on larger image datasets and using convolutional neural networks (we'll see these later).\n",
    "\n",
    "Let's check out more shapes of our data."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "fc4f768c-c3f6-454d-a633-673ad1d6eca0",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "fc4f768c-c3f6-454d-a633-673ad1d6eca0",
    "outputId": "fcac1ff4-5b9a-4459-a05e-77482f0e6776"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(60000, 60000, 10000, 10000)"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# How many samples are there? \n",
    "len(train_data.data), len(train_data.targets), len(test_data.data), len(test_data.targets)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6e0267d5-946b-4c53-af69-61acd3527972",
   "metadata": {
    "id": "6e0267d5-946b-4c53-af69-61acd3527972"
   },
   "source": [
    "So we've got 60,000 training samples and 10,000 testing samples.\n",
    "\n",
    "What classes are there?\n",
    "\n",
    "We can find these via the `.classes` attribute."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "e22849c6-d93f-4b38-8403-5ebf0deaf008",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "e22849c6-d93f-4b38-8403-5ebf0deaf008",
    "outputId": "6e18aa0f-b8a0-45ee-9f4e-8931bcdfbec0"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['T-shirt/top',\n",
       " 'Trouser',\n",
       " 'Pullover',\n",
       " 'Dress',\n",
       " 'Coat',\n",
       " 'Sandal',\n",
       " 'Shirt',\n",
       " 'Sneaker',\n",
       " 'Bag',\n",
       " 'Ankle boot']"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# See classes\n",
    "class_names = train_data.classes\n",
    "class_names"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "89483b15",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(torch.Size([1, 28, 28]), 9)"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "image.shape, label"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "abdd225c-5742-4d9c-8e8d-fb30a9c3cb6e",
   "metadata": {
    "id": "abdd225c-5742-4d9c-8e8d-fb30a9c3cb6e"
   },
   "source": [
    "Sweet! It looks like we're dealing with 10 different kinds of clothes.\n",
    "\n",
    "Because we're working with 10 different classes, it means our problem is **multi-class classification**.\n",
    "\n",
    "Let's get visual."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "fb625d80-6a98-471e-a758-4de0ce0f3a64",
   "metadata": {
    "id": "fb625d80-6a98-471e-a758-4de0ce0f3a64"
   },
   "source": [
    "### 1.2 Visualizing our data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "b1df1f2c-28c9-43bf-aaef-cf996c9ae1c5",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 469
    },
    "id": "b1df1f2c-28c9-43bf-aaef-cf996c9ae1c5",
    "outputId": "9bbdbb0d-eed3-408a-bd7b-03aa22cb35bb"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Image shape: torch.Size([1, 28, 28])\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "image, label = train_data[0]\n",
    "print(f\"Image shape: {image.shape}\")\n",
    "plt.imshow(image.squeeze()) # image shape is [1, 28, 28] (colour channels, height, width)\n",
    "plt.title(label);"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "adb19c5c-2f2b-4aaf-8300-256f3594e2db",
   "metadata": {
    "id": "adb19c5c-2f2b-4aaf-8300-256f3594e2db"
   },
   "source": [
    "We can turn the image into grayscale using the `cmap` parameter of `plt.imshow()`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "92f09917-88f7-4446-b65f-baae586914c9",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 452
    },
    "id": "92f09917-88f7-4446-b65f-baae586914c9",
    "outputId": "c702456b-607c-4214-8e03-4bd0b22b097f"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.imshow(image.squeeze(), cmap=\"gray\")\n",
    "plt.title(class_names[label]);"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9a09388a-d754-485f-aa26-4e7a0f782967",
   "metadata": {
    "id": "9a09388a-d754-485f-aa26-4e7a0f782967"
   },
   "source": [
    "Beautiful, well as beautiful as a pixelated grayscale ankle boot can get.\n",
    "\n",
    "Let's view a few more."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "7188ed7a-5959-48c4-ac7f-19129a2adc83",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 752
    },
    "id": "7188ed7a-5959-48c4-ac7f-19129a2adc83",
    "outputId": "98d50938-b984-4725-8949-d85bf3143555"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 900x900 with 16 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Plot more images\n",
    "torch.manual_seed(42)\n",
    "fig = plt.figure(figsize=(9, 9))\n",
    "rows, cols = 4, 4\n",
    "for i in range(1, rows * cols + 1):\n",
    "    random_idx = torch.randint(0, len(train_data), size=[1]).item()\n",
    "    img, label = train_data[random_idx]\n",
    "    fig.add_subplot(rows, cols, i)\n",
    "    plt.imshow(img.squeeze(), cmap=\"gray\")\n",
    "    plt.title(class_names[label])\n",
    "    plt.axis(False);"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f356fbe9-95b1-4f81-a82d-dc15b3adc06a",
   "metadata": {
    "id": "f356fbe9-95b1-4f81-a82d-dc15b3adc06a"
   },
   "source": [
    "Hmmm, this dataset doesn't look too aesthetic.\n",
    "\n",
    "But the principles we're going to learn on how to build a model for it will be similar across a wide range of computer vision problems.\n",
    "\n",
    "In essence, taking pixel values and building a model to find patterns in them to use on future pixel values.\n",
    "\n",
    "Plus, even for this small dataset (yes, even 60,000 images in deep learning is considered quite small), could you write a program to classify each one of them?\n",
    "\n",
    "You probably could.\n",
    "\n",
    "But I think coding a model in PyTorch would be faster.\n",
    "\n",
    "> **Question:** Do you think the above data can be model with only straight (linear) lines? Or do you think you'd also need non-straight (non-linear) lines?\n",
    "\n",
    "**Non-linearity required**"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "43cdd23d-bd1f-4e8c-ba20-22d2b6ac14b1",
   "metadata": {
    "id": "43cdd23d-bd1f-4e8c-ba20-22d2b6ac14b1"
   },
   "source": [
    "## 2. Prepare DataLoader\n",
    "\n",
    "Now we've got a dataset ready to go.\n",
    "\n",
    "The next step is to prepare it with a [`torch.utils.data.DataLoader`](https://pytorch.org/docs/stable/data.html#torch.utils.data.Dataset) or `DataLoader` for short.\n",
    "\n",
    "The `DataLoader` does what you think it might do.\n",
    "\n",
    "It helps load data into a model.\n",
    "\n",
    "For training and for inference.\n",
    "\n",
    "It turns a large `Dataset` into a Python iterable of smaller chunks.\n",
    "\n",
    "These smaller chunks are called **batches** or **mini-batches** and can be set by the `batch_size` parameter.\n",
    "\n",
    "Why do this?\n",
    "\n",
    "Because it's more computationally efficient.\n",
    "\n",
    "In an ideal world you could do the forward pass and backward pass across all of your data at once.\n",
    "\n",
    "But once you start using really large datasets, unless you've got infinite computing power, it's easier to break them up into batches.\n",
    "\n",
    "It also gives your model more opportunities to improve.\n",
    "\n",
    "With **mini-batches** (small portions of the data), gradient descent is performed more often per epoch (once per mini-batch rather than once per epoch).\n",
    "\n",
    "What's a good batch size?\n",
    "\n",
    "[32 is a good place to start](https://twitter.com/ylecun/status/989610208497360896?s=20&t=N96J_jotN--PYuJk2WcjMw) for a fair amount of problems.\n",
    "\n",
    "But since this is a value you can set (a **hyperparameter**) you can try all different kinds of values, though generally powers of 2 are used most often (e.g. 32, 64, 128, 256, 512).\n",
    "\n",
    "![an example of what a batched dataset looks like](https://raw.githubusercontent.com/mrdbourke/pytorch-deep-learning/main/images/03-batching-fashionmnist.png)\n",
    "*Batching FashionMNIST with a batch size of 32 and shuffle turned on. A similar batching process will occur for other datasets but will differ depending on the batch size.*\n",
    "\n",
    "Let's create `DataLoader`'s for our training and test sets. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "bb2dbf90-a326-43cb-b25b-71af142fafeb",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "bb2dbf90-a326-43cb-b25b-71af142fafeb",
    "outputId": "1f563408-3f50-4e8c-a15f-53e2f918b1ac"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Dataloaders: (<torch.utils.data.dataloader.DataLoader object at 0x000001DE5D271690>, <torch.utils.data.dataloader.DataLoader object at 0x000001DE56591330>)\n",
      "Length of train dataloader: 1875 batches of 32\n",
      "Length of test dataloader: 313 batches of 32\n"
     ]
    }
   ],
   "source": [
    "from torch.utils.data import DataLoader\n",
    "\n",
    "# Setup the batch size hyperparameter\n",
    "BATCH_SIZE = 32\n",
    "\n",
    "# Turn datasets into iterables (batches)\n",
    "train_dataloader = DataLoader(train_data, # dataset to turn into iterable\n",
    "    batch_size=BATCH_SIZE, # how many samples per batch? \n",
    "    shuffle=True # shuffle data every epoch?\n",
    ")\n",
    "\n",
    "test_dataloader = DataLoader(test_data,\n",
    "    batch_size=BATCH_SIZE,\n",
    "    shuffle=False # don't necessarily have to shuffle the testing data\n",
    ")\n",
    "\n",
    "# Let's check out what we've created\n",
    "print(f\"Dataloaders: {train_dataloader, test_dataloader}\") \n",
    "print(f\"Length of train dataloader: {len(train_dataloader)} batches of {BATCH_SIZE}\")\n",
    "print(f\"Length of test dataloader: {len(test_dataloader)} batches of {BATCH_SIZE}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "1656401b",
   "metadata": {},
   "outputs": [],
   "source": [
    "#loader_iter = iter(train_dataloader)\n",
    "#loader_iter"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "7a925ee7-484b-4149-be8f-3ad790172a5f",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "7a925ee7-484b-4149-be8f-3ad790172a5f",
    "outputId": "85815bd7-39e9-44ed-b974-9e30fff5b5ad"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(torch.Size([32, 1, 28, 28]), torch.Size([32]))"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Check out what's inside the training dataloader\n",
    "iterator = iter(train_dataloader)\n",
    "train_features_batch, train_labels_batch = next(iterator)\n",
    "train_features_batch.shape, train_labels_batch.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "07182f5f",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(torch.Size([32, 1, 28, 28]), torch.Size([32]))"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_features_batch2, train_labels_batch2 = next(iterator)\n",
    "train_features_batch2.shape, train_labels_batch2.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "36537044",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "60000"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(train_dataloader) * 32"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "d7139311",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n",
      "torch.Size([32, 1, 28, 28]) torch.Size([32])\n"
     ]
    }
   ],
   "source": [
    "for features, labels in train_dataloader: # loop over the batches\n",
    "    print(features.shape, labels.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4fee4cf8-ab73-4c81-8e5e-3c81691e799c",
   "metadata": {
    "id": "4fee4cf8-ab73-4c81-8e5e-3c81691e799c"
   },
   "source": [
    "And we can see that the data remains unchanged by checking a single sample. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "c863d66a-49be-43be-84dc-372a5d6fc2c2",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 463
    },
    "id": "c863d66a-49be-43be-84dc-372a5d6fc2c2",
    "outputId": "1052cbcb-6186-4dfe-b5f0-6968bde9fb21"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Image size: torch.Size([1, 28, 28])\n",
      "Label: 5, label size: torch.Size([])\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Show a sample\n",
    "#torch.manual_seed(42)\n",
    "random_idx = torch.randint(0, len(train_features_batch), size=[1]).item()\n",
    "img, label = train_features_batch[random_idx], train_labels_batch[random_idx]\n",
    "plt.imshow(img.squeeze(), cmap=\"gray\")\n",
    "plt.title(class_names[label])\n",
    "plt.axis(\"Off\");\n",
    "print(f\"Image size: {img.shape}\")\n",
    "print(f\"Label: {label}, label size: {label.shape}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "db1695cf-f53d-4c7c-ad39-dfed76533125",
   "metadata": {
    "id": "db1695cf-f53d-4c7c-ad39-dfed76533125"
   },
   "source": [
    "## 3. Model 0: Build a baseline model\n",
    "\n",
    "Data loaded and prepared!\n",
    "\n",
    "Time to build a **baseline model** by subclassing `nn.Module`.\n",
    "\n",
    "A **baseline model** is one of the simplest models you can imagine.\n",
    "\n",
    "You use the baseline as a starting point and try to improve upon it with subsequent, more complicated models.\n",
    "\n",
    "Our baseline will consist of two [`nn.Linear()`](https://pytorch.org/docs/stable/generated/torch.nn.Linear.html) layers.\n",
    "\n",
    "We've done this in a previous section but there's going to one slight difference.\n",
    "\n",
    "Because we're working with image data, we're going to use a different layer to start things off.\n",
    "\n",
    "And that's the [`nn.Flatten()`](https://pytorch.org/docs/stable/generated/torch.nn.Flatten.html) layer.\n",
    "\n",
    "`nn.Flatten()` compresses the dimensions of a tensor into a single vector.\n",
    "\n",
    "This is easier to understand when you see it."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "8d99bafd",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "torch.Size([1, 28, 28])"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_features_batch[0].shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "405319f1-f242-4bd9-90f5-3abdc50782ac",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "405319f1-f242-4bd9-90f5-3abdc50782ac",
    "outputId": "742cd0fe-c95f-4201-a469-f12733625784"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Shape before flattening: torch.Size([1, 28, 28]) -> [color_channels, height, width]\n",
      "Shape after flattening: torch.Size([1, 784]) -> [color_channels, height*width]\n"
     ]
    }
   ],
   "source": [
    "# Create a flatten layer\n",
    "flatten_model = nn.Flatten() # all nn modules function as a model (can do a forward pass)\n",
    "\n",
    "# Get a single sample\n",
    "x = train_features_batch[0]\n",
    "\n",
    "# Flatten the sample\n",
    "output = flatten_model(x) # perform forward pass\n",
    "\n",
    "# Print out what happened\n",
    "print(f\"Shape before flattening: {x.shape} -> [color_channels, height, width]\")\n",
    "print(f\"Shape after flattening: {output.shape} -> [color_channels, height*width]\")\n",
    "\n",
    "# Try uncommenting below and see what happens\n",
    "#print(x)\n",
    "#print(output)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "86bb7806-fca6-45af-8111-3e00e38f5be9",
   "metadata": {
    "id": "86bb7806-fca6-45af-8111-3e00e38f5be9"
   },
   "source": [
    "The `nn.Flatten()` layer took our shape from `[color_channels, height, width]` to `[color_channels, height*width]`.\n",
    "\n",
    "Why do this?\n",
    "\n",
    "Because we've now turned our pixel data from height and width dimensions into one long **feature vector**.\n",
    "\n",
    "And `nn.Linear()` layers like their inputs to be in the form of feature vectors.\n",
    "\n",
    "Let's create our first model using `nn.Flatten()` as the first layer. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "1449f427-6859-41ae-8133-50b58ffbce72",
   "metadata": {
    "id": "1449f427-6859-41ae-8133-50b58ffbce72"
   },
   "outputs": [],
   "source": [
    "from torch import nn\n",
    "class FashionMNISTModelV0(nn.Module):\n",
    "    def __init__(self, input_shape: int, hidden_units: int, output_shape: int):\n",
    "        super().__init__()\n",
    "        self.layer_stack = nn.Sequential(\n",
    "            nn.Flatten(), # neural networks like their inputs in vector form\n",
    "            nn.Linear(in_features=input_shape, out_features=hidden_units), # in_features = number of features in a data sample (784 pixels)\n",
    "            nn.Linear(in_features=hidden_units, out_features=output_shape)\n",
    "        )\n",
    "    \n",
    "    def forward(self, x):\n",
    "        return self.layer_stack(x)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4d1b50bf-d00b-485c-be00-b3e4de156fab",
   "metadata": {
    "id": "4d1b50bf-d00b-485c-be00-b3e4de156fab"
   },
   "source": [
    "Wonderful!\n",
    "\n",
    "We've got a baseline model class we can use, now let's instantiate a model.\n",
    "\n",
    "We'll need to set the following parameters:\n",
    "* `input_shape=784` - this is how many features you've got going in the model, in our case, it's one for every pixel in the target image (28 pixels high by 28 pixels wide = 784 features).\n",
    "* `hidden_units=10` - number of units/neurons in the hidden layer(s), this number could be whatever you want but to keep the model small we'll start with `10`.\n",
    "* `output_shape=len(class_names)` - since we're working with a multi-class classification problem, we need an output neuron per class in our dataset.\n",
    "\n",
    "Let's create an instance of our model and send to the CPU for now (we'll run a small test for running `model_0` on CPU vs. a similar model on GPU soon)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "dd18384a-76f9-4b5a-a013-fda077f16865",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "dd18384a-76f9-4b5a-a013-fda077f16865",
    "outputId": "e4b63839-d012-40db-a7f7-967a146fe566"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "FashionMNISTModelV0(\n",
       "  (layer_stack): Sequential(\n",
       "    (0): Flatten(start_dim=1, end_dim=-1)\n",
       "    (1): Linear(in_features=784, out_features=10, bias=True)\n",
       "    (2): Linear(in_features=10, out_features=10, bias=True)\n",
       "  )\n",
       ")"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "torch.manual_seed(42)\n",
    "\n",
    "# Need to setup model with input parameters\n",
    "model_0 = FashionMNISTModelV0(input_shape=train_features_batch[0].shape[1] * train_features_batch[0].shape[2], # 784, # one for every pixel (28x28)\n",
    "    hidden_units=10, # how many units in the hiden layer\n",
    "    output_shape=len(class_names) # one for every class\n",
    ")\n",
    "model_0.to(\"cpu\") # keep model on CPU to begin with "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "271b9e23",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tensor([[-0.0315,  0.3171,  0.0531, -0.2525,  0.5959,  0.2112,  0.3233,  0.2694,\n",
       "         -0.1004,  0.0157]], grad_fn=<AddmmBackward0>)"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dummy_x = torch.rand([1,1,28,28])\n",
    "model_0(dummy_x)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "03243179-1cdc-45d9-8b8c-82538ac02e9c",
   "metadata": {
    "id": "03243179-1cdc-45d9-8b8c-82538ac02e9c"
   },
   "source": [
    "### 3.1 Setup loss, optimizer and evaluation metrics\n",
    "\n",
    "Since we're working on a classification problem, let's bring in our [`helper_functions.py` script](https://github.com/mrdbourke/pytorch-deep-learning/blob/main/helper_functions.py) and subsequently the `accuracy_fn()` we defined in [notebook 02](https://www.learnpytorch.io/02_pytorch_classification/).\n",
    "\n",
    "> **Note:** Rather than importing and using our own accuracy function or evaluation metric(s), you could import various evaluation metrics from the [TorchMetrics package](https://torchmetrics.readthedocs.io/en/latest/)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "31c91f17-d810-46a4-97c3-c734f93430b1",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "31c91f17-d810-46a4-97c3-c734f93430b1",
    "outputId": "d2333811-f5fa-426f-90a7-c884fe2493df"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "helper_functions.py already exists, skipping download\n"
     ]
    }
   ],
   "source": [
    "import requests\n",
    "from pathlib import Path \n",
    "\n",
    "# Download helper functions from Learn PyTorch repo (if not already downloaded)\n",
    "if Path(\"helper_functions.py\").is_file():\n",
    "  print(\"helper_functions.py already exists, skipping download\")\n",
    "else:\n",
    "  print(\"Downloading helper_functions.py\")\n",
    "  # Note: you need the \"raw\" GitHub URL for this to work\n",
    "  request = requests.get(\"https://raw.githubusercontent.com/mrdbourke/pytorch-deep-learning/main/helper_functions.py\")\n",
    "  with open(\"helper_functions.py\", \"wb\") as f:\n",
    "    f.write(request.content)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "f27112c0",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "model_evaluation.py already exsits, skipping download\n"
     ]
    }
   ],
   "source": [
    "if Path(\"model_evaluation.py\").is_file():\n",
    "    print(\"model_evaluation.py already exsits, skipping download\")\n",
    "else:\n",
    "    print(\"Downloading model_evaluation.py\")\n",
    "    request = requests.get(\"https://github.com/sergio-sanz-rodriguez/custom_python_functions/blob/main/model_evaluation.py\")\n",
    "    with open(\"model_evaluation.py\", \"wb\") as f:\n",
    "        f.write(request.content)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "ce3d13b8-f018-4b44-8bba-375074dc4c5f",
   "metadata": {
    "id": "ce3d13b8-f018-4b44-8bba-375074dc4c5f"
   },
   "outputs": [],
   "source": [
    "# Import accuracy metric\n",
    "from helper_functions import accuracy_fn # Note: could also use torchmetrics.Accuracy(task = 'multiclass', num_classes=len(class_names)).to(device)\n",
    "\n",
    "# Setup loss function and optimizer\n",
    "loss_fn = nn.CrossEntropyLoss() # this is also called \"criterion\"/\"cost function\" in some places\n",
    "optimizer = torch.optim.SGD(params=model_0.parameters(), lr=0.1)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4109f867-83f2-4394-a925-8acdc63ccffe",
   "metadata": {
    "id": "4109f867-83f2-4394-a925-8acdc63ccffe"
   },
   "source": [
    "### 3.2 Creating a function to time our experiments\n",
    "\n",
    "Loss function and optimizer ready!\n",
    "\n",
    "It's time to start training a model.\n",
    "\n",
    "But how about we do a little experiment while we train.\n",
    "\n",
    "I mean, let's make a timing function to measure the time it takes our model to train on CPU versus using a GPU.\n",
    "\n",
    "We'll train this model on the CPU but the next one on the GPU and see what happens.\n",
    "\n",
    "Our timing function will import the [`timeit.default_timer()` function](https://docs.python.org/3/library/timeit.html#timeit.default_timer) from the Python [`timeit` module](https://docs.python.org/3/library/timeit.html)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "31adc3fe-ce90-4b4e-b0d4-3613abae5714",
   "metadata": {
    "id": "31adc3fe-ce90-4b4e-b0d4-3613abae5714"
   },
   "outputs": [],
   "source": [
    "from timeit import default_timer as timer \n",
    "def print_train_time(start: float, end: float, device: torch.device = None):\n",
    "    \"\"\"Prints difference between start and end time.\n",
    "\n",
    "    Args:\n",
    "        start (float): Start time of computation (preferred in timeit format). \n",
    "        end (float): End time of computation.\n",
    "        device ([type], optional): Device that compute is running on. Defaults to None.\n",
    "\n",
    "    Returns:\n",
    "        float: time between start and end in seconds (higher is longer).\n",
    "    \"\"\"\n",
    "    total_time = end - start\n",
    "    print(f\"Train time on {device}: {total_time:.3f} seconds\")\n",
    "    return total_time"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "07b9560e-f5dc-45d6-b3b2-ddae17a71b34",
   "metadata": {
    "id": "07b9560e-f5dc-45d6-b3b2-ddae17a71b34"
   },
   "source": [
    "### 3.3 Creating a training loop and training a model on batches of data\n",
    "\n",
    "Beautiful!\n",
    "\n",
    "Looks like we've got all of the pieces of the puzzle ready to go, a timer, a loss function, an optimizer, a model and most importantly, some data.\n",
    "\n",
    "Let's now create a training loop and a testing loop to train and evaluate our model.\n",
    "\n",
    "We'll be using the same steps as the previous notebook(s), though since our data is now in batch form, we'll add another loop to loop through our data batches.\n",
    "\n",
    "Our data batches are contained within our `DataLoader`s, `train_dataloader` and `test_dataloader` for the training and test data splits respectively.\n",
    "\n",
    "A batch is `BATCH_SIZE` samples of `X` (features) and `y` (labels), since we're using `BATCH_SIZE=32`, our batches have 32 samples of images and targets.\n",
    "\n",
    "And since we're computing on batches of data, our loss and evaluation metrics will be calculated **per batch** rather than across the whole dataset.\n",
    "\n",
    "This means we'll have to divide our loss and accuracy values by the number of batches in each dataset's respective dataloader. \n",
    "\n",
    "Let's step through it: \n",
    "1. Loop through epochs.\n",
    "2. Loop through training batches, perform training steps, calculate the train loss *per batch*.\n",
    "3. Loop through testing batches, perform testing steps, calculate the test loss *per batch*.\n",
    "4. Print out what's happening.\n",
    "5. Time it all (for fun).\n",
    "\n",
    "A fair few steps but...\n",
    "\n",
    "...if in doubt, code it out. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "8833decb",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "torch.Size([10, 784])\n",
      "torch.Size([10])\n",
      "torch.Size([10, 10])\n",
      "torch.Size([10])\n"
     ]
    }
   ],
   "source": [
    "for param in model_0.parameters():\n",
    "    print(param.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "49ed7d43",
   "metadata": {},
   "source": [
    "torch.Size([10, 784]): This represents the weights connecting the 784 input features to the 10 hidden units. There are 10 neurons (one per row), each with 784 weights (one per column).  \n",
    "torch.Size([10]): Each of the 10 hidden neurons also has an associated bias term, so there are 10 bias values.  \n",
    "torch.Size([10, 10]): This represents the weights connecting the 10 hidden units to the 10 output neurons.  \n",
    "torch.Size([10]): There are 10 bias terms, one for each of the 10 output neurons.  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "id": "0ff68189",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Total number of parameters: 7960\n"
     ]
    }
   ],
   "source": [
    "# Calculate total number of parameters in the model\n",
    "def calculate_total_params(model):\n",
    "    total_params = sum(p.numel() for p in model.parameters())\n",
    "    return total_params\n",
    "\n",
    "total_params = calculate_total_params(model_0)\n",
    "print(f\"Total number of parameters: {total_params}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "id": "c07bbf10-81e3-47f0-990d-9a4a838276ab",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 587,
     "referenced_widgets": [
      "0bd8f8b5ff4d4b50b03e3a65cc1446f0",
      "430d171cfd584196ad0fa3e1cd0a286c",
      "618fb3cf63a94da9ad5f29a3d9a87ac3",
      "3524e24faad44aa38926b40b2d590f6b",
      "c01ca4def9224135ad367b6f8dbbae62",
      "decc1966e6a84973839efc0c65f51790",
      "39fc424b6cef4e98a80a342f530be99b",
      "e929063168354b018bbf0bb45fdfef1f",
      "d62646457b284fcb8aeac382b77eb942",
      "5c0883aa74f94568850741dad118cb88",
      "e44697d8dd41492e8619a860b3911e19"
     ]
    },
    "id": "c07bbf10-81e3-47f0-990d-9a4a838276ab",
    "outputId": "3fb70da8-1a65-42bb-a684-85f0d1dd11c0"
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\Users\\ssre_\\.conda\\envs\\.venv_pytorch_gpu\\lib\\site-packages\\tqdm\\auto.py:21: TqdmWarning: IProgress not found. Please update jupyter and ipywidgets. See https://ipywidgets.readthedocs.io/en/stable/user_install.html\n",
      "  from .autonotebook import tqdm as notebook_tqdm\n",
      "  0%|          | 0/3 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch: 0\n",
      "-------\n",
      "Looked at 0/60000 samples\n",
      "Looked at 12800/60000 samples\n",
      "Looked at 25600/60000 samples\n",
      "Looked at 38400/60000 samples\n",
      "Looked at 51200/60000 samples\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      " 33%|███▎      | 1/3 [00:18<00:36, 18.18s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Train loss: 0.56665 | Test loss: 0.50300, Test acc: 81.91%\n",
      "\n",
      "Epoch: 1\n",
      "-------\n",
      "Looked at 0/60000 samples\n",
      "Looked at 12800/60000 samples\n",
      "Looked at 25600/60000 samples\n",
      "Looked at 38400/60000 samples\n",
      "Looked at 51200/60000 samples\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      " 67%|██████▋   | 2/3 [00:36<00:17, 17.99s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Train loss: 0.49440 | Test loss: 0.49818, Test acc: 82.85%\n",
      "\n",
      "Epoch: 2\n",
      "-------\n",
      "Looked at 0/60000 samples\n",
      "Looked at 12800/60000 samples\n",
      "Looked at 25600/60000 samples\n",
      "Looked at 38400/60000 samples\n",
      "Looked at 51200/60000 samples\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 3/3 [00:54<00:00, 18.13s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Train loss: 0.47912 | Test loss: 0.50100, Test acc: 83.13%\n",
      "\n",
      "Train time on cpu: 54.386 seconds\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "# Import tqdm for progress bar\n",
    "#!pip install tqdm\n",
    "from tqdm.auto import tqdm # prints the training progress\n",
    "\n",
    "# Set the seed and start the timer\n",
    "torch.manual_seed(42)\n",
    "train_time_start_on_cpu = timer()\n",
    "\n",
    "# Set the number of epochs (we'll keep this small for faster training times)\n",
    "epochs = 3\n",
    "\n",
    "# Create training and testing loop\n",
    "for epoch in tqdm(range(epochs)):\n",
    "    print(f\"Epoch: {epoch}\\n-------\")\n",
    "    ### Training\n",
    "    train_loss = 0\n",
    "    # Add a loop to loop through training batches\n",
    "    for batch, (X, y) in enumerate(train_dataloader): #Important for loop to learn\n",
    "        model_0.train() \n",
    "        # 1. Forward pass\n",
    "        y_pred = model_0(X)\n",
    "\n",
    "        # 2. Calculate loss (per batch)\n",
    "        loss = loss_fn(y_pred, y)\n",
    "        train_loss += loss # accumulatively add up the loss per epoch \n",
    "\n",
    "        # 3. Optimizer zero grad\n",
    "        optimizer.zero_grad()\n",
    "\n",
    "        # 4. Loss backward\n",
    "        loss.backward()\n",
    "\n",
    "        # 5. Optimizer step\n",
    "        optimizer.step()\n",
    "\n",
    "        # Print out how many samples have been seen\n",
    "        if batch % 400 == 0:\n",
    "            print(f\"Looked at {batch * len(X)}/{len(train_dataloader.dataset)} samples\")\n",
    "\n",
    "    # Divide total train loss by length of train dataloader (average loss per batch per epoch)\n",
    "    train_loss /= len(train_dataloader)\n",
    "    \n",
    "    ### Testing\n",
    "    # Setup variables for accumulatively adding up loss and accuracy \n",
    "    test_loss, test_acc = 0, 0 \n",
    "    model_0.eval()\n",
    "    with torch.inference_mode():\n",
    "        for X, y in test_dataloader:\n",
    "            # 1. Forward pass\n",
    "            test_pred = model_0(X)\n",
    "           \n",
    "            # 2. Calculate loss (accumatively)\n",
    "            test_loss += loss_fn(test_pred, y) # accumulatively add up the loss per epoch\n",
    "            \n",
    "            # 3. Calculate accuracy (preds need to be same as y_true)\n",
    "            test_acc += accuracy_fn(y_true=y, y_pred=test_pred.argmax(dim=1))\n",
    "            #test_acc += accuracy_fn(y_true=y, y_pred=torch.softmax(test_pred, dim=1).argmax(dim=1))            \n",
    "        \n",
    "        # Calculations on test metrics need to happen inside torch.inference_mode()\n",
    "        # Divide total test loss by length of test dataloader (per batch)\n",
    "        test_loss /= len(test_dataloader)\n",
    "\n",
    "        # Divide total accuracy by length of test dataloader (per batch)\n",
    "        test_acc /= len(test_dataloader)\n",
    "\n",
    "    ## Print out what's happening\n",
    "    print(f\"\\nTrain loss: {train_loss:.5f} | Test loss: {test_loss:.5f}, Test acc: {test_acc:.2f}%\\n\")\n",
    "\n",
    "# Calculate training time      \n",
    "train_time_end_on_cpu = timer()\n",
    "total_train_time_model_0 = print_train_time(start=train_time_start_on_cpu, \n",
    "                                           end=train_time_end_on_cpu,\n",
    "                                           device=str(next(model_0.parameters()).device))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7b02a939-a3a1-4a9d-bb9d-62928def2ded",
   "metadata": {
    "id": "7b02a939-a3a1-4a9d-bb9d-62928def2ded"
   },
   "source": [
    "Nice! Looks like our baseline model did fairly well.\n",
    "\n",
    "It didn't take too long to train either, even just on the CPU, I wonder if it'll speed up on the GPU?\n",
    "\n",
    "Let's write some code to evaluate our model."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7442511b-bfe9-4ec7-9f5b-9c808f8e560b",
   "metadata": {
    "id": "7442511b-bfe9-4ec7-9f5b-9c808f8e560b"
   },
   "source": [
    "## 4. Make predictions and get Model 0 results\n",
    "\n",
    "Since we're going to be building a few models, it's a good idea to write some code to evaluate them all in similar ways.\n",
    "\n",
    "Namely, let's create a function that takes in a trained model, a `DataLoader`, a loss function and an accuracy function.\n",
    "\n",
    "The function will use the model to make predictions on the data in the `DataLoader` and then we can evaluate those predictions using the loss function and accuracy function."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "id": "8317dd04-9de2-4fd7-97bd-1e202621397d",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "8317dd04-9de2-4fd7-97bd-1e202621397d",
    "outputId": "7132105e-6f51-43d3-efa7-f0231a21addc"
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 313/313 [00:02<00:00, 131.13it/s]\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "{'model_name': 'FashionMNISTModelV0',\n",
       " 'model_loss': 0.5010038018226624,\n",
       " 'model_acc': 83.12699680511182}"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "torch.manual_seed(42)\n",
    "def eval_model(model: torch.nn.Module, \n",
    "               data_loader: torch.utils.data.DataLoader, \n",
    "               loss_fn: torch.nn.Module, \n",
    "               accuracy_fn):\n",
    "    \"\"\"Returns a dictionary containing the results of model predicting on data_loader.\n",
    "\n",
    "    Args:\n",
    "        model (torch.nn.Module): A PyTorch model capable of making predictions on data_loader.\n",
    "        data_loader (torch.utils.data.DataLoader): The target dataset to predict on.\n",
    "        loss_fn (torch.nn.Module): The loss function of model.\n",
    "        accuracy_fn: An accuracy function to compare the models predictions to the truth labels.\n",
    "\n",
    "    Returns:\n",
    "        (dict): Results of model making predictions on data_loader.\n",
    "    \"\"\"\n",
    "    loss, acc = 0, 0\n",
    "    model.eval()\n",
    "    with torch.inference_mode():\n",
    "        for X, y in tqdm(data_loader):\n",
    "            # Make predictions with the model\n",
    "            y_pred = model(X)\n",
    "            \n",
    "            # Accumulate the loss and accuracy values per batch\n",
    "            loss += loss_fn(y_pred, y)\n",
    "            acc += accuracy_fn(y_true=y, \n",
    "                                y_pred=y_pred.argmax(dim=1)) # For accuracy, need the prediction labels (logits -> pred_prob -> pred_labels)\n",
    "        \n",
    "        # Scale loss and acc to find the average loss/acc per batch\n",
    "        loss /= len(data_loader)\n",
    "        acc /= len(data_loader)\n",
    "        \n",
    "    return {\"model_name\": model.__class__.__name__, # only works when model was created with a class\n",
    "            \"model_loss\": loss.item(),\n",
    "            \"model_acc\": acc}\n",
    "\n",
    "# Calculate model 0 results on test dataset\n",
    "model_0_results = eval_model(model=model_0, data_loader=test_dataloader,\n",
    "    loss_fn=loss_fn, accuracy_fn=accuracy_fn\n",
    ")\n",
    "model_0_results"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a39c3042-1262-4d1f-b33e-c8e2ba6781d3",
   "metadata": {
    "id": "a39c3042-1262-4d1f-b33e-c8e2ba6781d3"
   },
   "source": [
    "Looking good!\n",
    "\n",
    "We can use this dictionary to compare the baseline model results to other models later on."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3b76784d-4cdb-43d2-a6da-8e4da9a812a9",
   "metadata": {
    "id": "3b76784d-4cdb-43d2-a6da-8e4da9a812a9"
   },
   "source": [
    "## 5. Setup device agnostic-code (for using a GPU if there is one)\n",
    "We've seen how long it takes to train ma PyTorch model on 60,000 samples on CPU.\n",
    "\n",
    "> **Note:** Model training time is dependent on hardware used. Generally, more processors means faster training and smaller models on smaller datasets will often train faster than large models and large datasets.\n",
    "\n",
    "Now let's setup some [device-agnostic code](https://pytorch.org/docs/stable/notes/cuda.html#best-practices) for our models and data to run on GPU if it's available.\n",
    "\n",
    "If you're running this notebook on Google Colab, and you don't a GPU turned on yet, it's now time to turn one on via `Runtime -> Change runtime type -> Hardware accelerator -> GPU`. If you do this, your runtime will likely reset and you'll have to run all of the cells above by going `Runtime -> Run before`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "id": "17b69fe9-f974-4538-922c-20c5cc8220cc",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 35
    },
    "id": "17b69fe9-f974-4538-922c-20c5cc8220cc",
    "outputId": "10c3b74b-4db7-4a30-8c3a-5a259d1f54b8"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'cuda'"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Setup device agnostic code\n",
    "import torch\n",
    "device = \"cuda\" if torch.cuda.is_available() else \"cpu\"\n",
    "device"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "514021a8-d6f2-47f3-ab50-55f844e42310",
   "metadata": {
    "id": "514021a8-d6f2-47f3-ab50-55f844e42310"
   },
   "source": [
    "Beautiful!\n",
    "\n",
    "Let's build another model."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d7893907-5f82-4c5e-8fde-fa542a9f25af",
   "metadata": {
    "id": "d7893907-5f82-4c5e-8fde-fa542a9f25af"
   },
   "source": [
    "## 6. Model 1: Building a better model with non-linearity\n",
    "\n",
    "We learned about [the power of non-linearity in notebook 02](https://www.learnpytorch.io/02_pytorch_classification/#6-the-missing-piece-non-linearity).\n",
    "\n",
    "Seeing the data we've been working with, do you think it needs non-linear functions?\n",
    "\n",
    "And remember, linear means straight and non-linear means non-straight.\n",
    "\n",
    "Let's find out.\n",
    "\n",
    "We'll do so by recreating a similar model to before, except this time we'll put non-linear functions (`nn.ReLU()`) in between each linear layer."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "id": "2ccce5f2-b1e5-47a6-a7f3-6bc096b35ffb",
   "metadata": {
    "id": "2ccce5f2-b1e5-47a6-a7f3-6bc096b35ffb"
   },
   "outputs": [],
   "source": [
    "# Create a model with non-linear and linear layers\n",
    "class FashionMNISTModelV1(nn.Module):\n",
    "    def __init__(self, input_shape: int, hidden_units: int, output_shape: int):\n",
    "        super().__init__()\n",
    "        self.layer_stack = nn.Sequential(\n",
    "            nn.Flatten(), # flatten inputs into single vector\n",
    "            nn.Linear(in_features=input_shape, out_features=hidden_units),\n",
    "            nn.ReLU(),\n",
    "            nn.Linear(in_features=hidden_units, out_features=output_shape),\n",
    "            nn.ReLU()\n",
    "        )\n",
    "    \n",
    "    def forward(self, x: torch.Tensor):\n",
    "        return self.layer_stack(x)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4b4b7a2f-4834-4aa1-a8e2-b6e3e2b49224",
   "metadata": {
    "id": "4b4b7a2f-4834-4aa1-a8e2-b6e3e2b49224"
   },
   "source": [
    "That looks good.\n",
    "\n",
    "Now let's instantiate it with the same settings we used before.\n",
    "\n",
    "We'll need `input_shape=784` (equal to the number of features of our image data), `hidden_units=10` (starting small and the same as our baseline model) and `output_shape=len(class_names)` (one output unit per class).\n",
    "\n",
    "> **Note:** Notice how we kept most of the settings of our model the same except for one change: adding non-linear layers. This is a standard practice for running a series of machine learning experiments, change one thing and see what happens, then do it again, again, again."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "id": "907091ec-7e46-470b-a305-788a3009b837",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "907091ec-7e46-470b-a305-788a3009b837",
    "outputId": "4cecd2df-2918-4368-fa33-7aea8f958d8f"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "device(type='cuda', index=0)"
      ]
     },
     "execution_count": 66,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "torch.manual_seed(42)\n",
    "model_1 = FashionMNISTModelV1(input_shape=784, # number of input features\n",
    "    hidden_units=10,\n",
    "    output_shape=len(class_names) # number of output classes desired\n",
    ").to(device) # send model to GPU if it's available\n",
    "next(model_1.parameters()).device # check model device"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "id": "799e8463",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Total number of parameters: 7960\n"
     ]
    }
   ],
   "source": [
    "total_params = calculate_total_params(model_1)\n",
    "print(f\"Total number of parameters: {total_params}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b54a4e9d-a7ad-404c-920f-485fcff18a92",
   "metadata": {
    "id": "b54a4e9d-a7ad-404c-920f-485fcff18a92"
   },
   "source": [
    "### 6.1 Setup loss, optimizer and evaluation metrics\n",
    "\n",
    "As usual, we'll setup a loss function, an optimizer and an evaluation metric (we could do multiple evaluation metrics but we'll stick with accuracy for now)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 96,
   "id": "fe7e463b-d46c-4f00-853c-fdf0a28d74c8",
   "metadata": {
    "id": "fe7e463b-d46c-4f00-853c-fdf0a28d74c8"
   },
   "outputs": [],
   "source": [
    "from helper_functions import accuracy_fn\n",
    "loss_fn = nn.CrossEntropyLoss()\n",
    "optimizer = torch.optim.SGD(params=model_1.parameters(), \n",
    "                            lr=0.1)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1eb30af6-a355-49a2-a59f-25169fd27a6e",
   "metadata": {
    "id": "1eb30af6-a355-49a2-a59f-25169fd27a6e"
   },
   "source": [
    "### 6.2 Functionizing training and test loops\n",
    "\n",
    "So far we've been writing train and test loops over and over. \n",
    "\n",
    "Let's write them again but this time we'll put them in functions so they can be called again and again.\n",
    "\n",
    "And because we're using device-agnostic code now, we'll be sure to call `.to(device)` on our feature (`X`) and target (`y`) tensors.\n",
    "\n",
    "For the training loop we'll create a function called `train_step()` which takes in a model, a `DataLoader` a loss function and an optimizer.\n",
    "\n",
    "The testing loop will be similar but it'll be called `test_step()` and it'll take in a model, a `DataLoader`, a loss function and an evaluation function.\n",
    "\n",
    "> **Note:** Since these are functions, you can customize them in any way you like. What we're making here can be considered barebones training and testing functions for our specific classification use case."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 97,
   "id": "3d239ed2-4028-4603-8db3-ffca2b727819",
   "metadata": {
    "id": "3d239ed2-4028-4603-8db3-ffca2b727819"
   },
   "outputs": [],
   "source": [
    "def train_step(model: torch.nn.Module,\n",
    "               data_loader: torch.utils.data.DataLoader,\n",
    "               loss_fn: torch.nn.Module,\n",
    "               optimizer: torch.optim.Optimizer,\n",
    "               accuracy_fn,\n",
    "               device: torch.device = device):\n",
    "    train_loss, train_acc = 0, 0\n",
    "    model.to(device)\n",
    "    for batch, (X, y) in enumerate(data_loader):\n",
    "        # Send data to GPU\n",
    "        X, y = X.to(device), y.to(device)\n",
    "\n",
    "        # 1. Forward pass\n",
    "        y_pred = model(X)\n",
    "\n",
    "        # 2. Calculate loss\n",
    "        loss = loss_fn(y_pred, y)\n",
    "        train_loss += loss\n",
    "        train_acc += accuracy_fn(y_true=y,\n",
    "                                 y_pred=y_pred.argmax(dim=1)) # Go from logits -> pred labels\n",
    "\n",
    "        # 3. Optimizer zero grad\n",
    "        optimizer.zero_grad()\n",
    "\n",
    "        # 4. Loss backward\n",
    "        loss.backward()\n",
    "\n",
    "        # 5. Optimizer step\n",
    "        optimizer.step()\n",
    "\n",
    "    # Calculate loss and accuracy per epoch and print out what's happening\n",
    "    train_loss /= len(data_loader)\n",
    "    train_acc /= len(data_loader)\n",
    "    print(f\"Train loss: {train_loss:.5f} | Train accuracy: {train_acc:.2f}%\")\n",
    "\n",
    "def test_step(data_loader: torch.utils.data.DataLoader,\n",
    "              model: torch.nn.Module,\n",
    "              loss_fn: torch.nn.Module,\n",
    "              accuracy_fn,\n",
    "              device: torch.device = device):\n",
    "    test_loss, test_acc = 0, 0\n",
    "    model.to(device)\n",
    "    model.eval() # put model in eval mode\n",
    "    # Turn on inference context manager\n",
    "    with torch.inference_mode(): \n",
    "        for X, y in data_loader:\n",
    "            # Send data to GPU\n",
    "            X, y = X.to(device), y.to(device)\n",
    "            \n",
    "            # 1. Forward pass\n",
    "            test_pred = model(X)\n",
    "            \n",
    "            # 2. Calculate loss and accuracy\n",
    "            test_loss += loss_fn(test_pred, y)\n",
    "            test_acc += accuracy_fn(y_true=y,\n",
    "                y_pred=test_pred.argmax(dim=1) # Go from logits -> pred labels\n",
    "            )\n",
    "        \n",
    "        # Adjust metrics and print out\n",
    "        test_loss /= len(data_loader)\n",
    "        test_acc /= len(data_loader)\n",
    "        print(f\"Test loss: {test_loss:.5f} | Test accuracy: {test_acc:.2f}%\\n\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e44121b6-c4be-4909-9175-dc9bd8dc6273",
   "metadata": {
    "id": "e44121b6-c4be-4909-9175-dc9bd8dc6273"
   },
   "source": [
    "Woohoo!\n",
    "\n",
    "Now we've got some functions for training and testing our model, let's run them.\n",
    "\n",
    "We'll do so inside another loop for each epoch.\n",
    "\n",
    "That way for each epoch we're going a training and a testing step.\n",
    "\n",
    "> **Note:** You can customize how often you do a testing step. Sometimes people do them every five epochs or 10 epochs or in our case, every epoch.\n",
    "\n",
    "Let's also time things to see how long our code takes to run on the GPU."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 98,
   "id": "2bb8094b-01a0-4b84-9526-ba8888d04901",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 327,
     "referenced_widgets": [
      "3ee8f4a32dae40a2954869aa28d511af",
      "9bdbfed6aaa64648ac9624541a719134",
      "a7e31e6725a0417495bb5d8d9bb0eedb",
      "8a07bf3a83cf44b09ebec23372699dd4",
      "4da7f6dcecfc44928a784709a2f85c67",
      "85241944b82749bda4b5b6ff50f484b2",
      "b139c87d10be44229d2f65d356912c25",
      "b684374f8a3c41cb887142dd2c4a0c94",
      "325e5b7b95db4289b3ee1bd6dbfc4a6c",
      "987db9e4bab746ff9d393aa1409cf628",
      "dd5dcc8d0c424965ba5a329efbf725cc"
     ]
    },
    "id": "2bb8094b-01a0-4b84-9526-ba8888d04901",
    "outputId": "83769d2d-6f3b-4704-e443-cfc4ef52cc81"
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "  0%|          | 0/3 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch: 0\n",
      "---------\n",
      "Train loss: 0.55983 | Train accuracy: 79.64%\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      " 33%|███▎      | 1/3 [00:19<00:38, 19.22s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Test loss: 0.49927 | Test accuracy: 81.20%\n",
      "\n",
      "Epoch: 1\n",
      "---------\n",
      "Train loss: 0.46043 | Train accuracy: 83.23%\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      " 67%|██████▋   | 2/3 [00:39<00:19, 19.65s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Test loss: 0.47899 | Test accuracy: 82.76%\n",
      "\n",
      "Epoch: 2\n",
      "---------\n",
      "Train loss: 0.43992 | Train accuracy: 84.05%\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 3/3 [00:58<00:00, 19.65s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Test loss: 0.46016 | Test accuracy: 84.20%\n",
      "\n",
      "Train time on cuda: 58.957 seconds\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "torch.manual_seed(42)\n",
    "\n",
    "# Measure time\n",
    "from timeit import default_timer as timer\n",
    "train_time_start_on_gpu = timer()\n",
    "\n",
    "epochs = 3\n",
    "for epoch in tqdm(range(epochs)):\n",
    "    print(f\"Epoch: {epoch}\\n---------\")\n",
    "    train_step(data_loader=train_dataloader, \n",
    "        model=model_1, \n",
    "        loss_fn=loss_fn,\n",
    "        optimizer=optimizer,\n",
    "        accuracy_fn=accuracy_fn\n",
    "    )\n",
    "    test_step(data_loader=test_dataloader,\n",
    "        model=model_1,\n",
    "        loss_fn=loss_fn,\n",
    "        accuracy_fn=accuracy_fn\n",
    "    )\n",
    "\n",
    "train_time_end_on_gpu = timer()\n",
    "total_train_time_model_1 = print_train_time(start=train_time_start_on_gpu,\n",
    "                                            end=train_time_end_on_gpu,\n",
    "                                            device=device)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "719b8eb9-9a7f-42ed-a49f-5eedc6fdd720",
   "metadata": {
    "id": "719b8eb9-9a7f-42ed-a49f-5eedc6fdd720"
   },
   "source": [
    "Excellent!\n",
    "\n",
    "Our model trained but the training time took longer?\n",
    "\n",
    "> **Note:** The training time on CUDA vs CPU will depend largely on the quality of the CPU/GPU you're using. Read on for a more explained answer.\n",
    "\n",
    "> **Question:** \"I used a a GPU but my model didn't train faster, why might that be?\"\n",
    ">\n",
    "> **Answer:** Well, one reason could be because your dataset and model are both so small (like the dataset and model we're working with) the benefits of using a GPU are outweighed by the time it actually takes to transfer the data there.\n",
    "> \n",
    "> There's a small bottleneck between copying data from the CPU memory (default) to the GPU memory.\n",
    ">\n",
    "> So for smaller models and datasets, the CPU might actually be the optimal place to compute on.\n",
    ">\n",
    "> But for larger datasets and models, the speed of computing the GPU can offer usually far outweighs the cost of getting the data there.\n",
    ">\n",
    "> However, this is largely dependant on the hardware you're using. With practice, you will get used to where the best place to train your models is. \n",
    "\n",
    "Let's evaluate our trained `model_1` using our `eval_model()` function and see how it went."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 99,
   "id": "32a544e3-9dbe-4aa1-b074-22e28b8f2f2a",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 381
    },
    "id": "32a544e3-9dbe-4aa1-b074-22e28b8f2f2a",
    "outputId": "bab29648-1e35-4f01-9efe-fa4d2030cddb"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'model_name': 'FashionMNISTModelV1',\n",
       " 'model_loss': 0.4601558446884155,\n",
       " 'model_acc': 84.1952875399361}"
      ]
     },
     "execution_count": 99,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "torch.manual_seed(42)\n",
    "\n",
    "# Note: This will error due to `eval_model()` not using device agnostic code \n",
    "model_1_results = eval_model(model=model_1, \n",
    "    data_loader=test_dataloader,\n",
    "    loss_fn=loss_fn, \n",
    "    accuracy_fn=accuracy_fn) \n",
    "model_1_results "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6a3481a5-489d-4db9-ac95-c3ce385978b7",
   "metadata": {
    "id": "6a3481a5-489d-4db9-ac95-c3ce385978b7"
   },
   "source": [
    "Oh no! \n",
    "\n",
    "It looks like our `eval_model()` function errors out with:\n",
    "\n",
    "> `RuntimeError: Expected all tensors to be on the same device, but found at least two devices, cuda:0 and cpu! (when checking argument for argument mat1 in method wrapper_addmm)`\n",
    "\n",
    "It's because we've setup our data and model to use device-agnostic code but not our evaluation function.\n",
    "\n",
    "How about we fix that by passing a target `device` parameter to our `eval_model()` function?\n",
    "\n",
    "Then we'll try calculating the results again."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 94,
   "id": "f3665d99-1adc-4d9f-bfc6-e5601a80691c",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "f3665d99-1adc-4d9f-bfc6-e5601a80691c",
    "outputId": "05312922-d30b-4c09-9989-963a4a579bf8"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'model_name': 'FashionMNISTModelV1',\n",
       " 'model_loss': 2.30334734916687,\n",
       " 'model_acc': 9.804313099041533}"
      ]
     },
     "execution_count": 94,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Move values to device\n",
    "torch.manual_seed(42)\n",
    "def eval_model(model: torch.nn.Module, \n",
    "               data_loader: torch.utils.data.DataLoader, \n",
    "               loss_fn: torch.nn.Module, \n",
    "               accuracy_fn, \n",
    "               device: torch.device = device):\n",
    "    \"\"\"Evaluates a given model on a given dataset.\n",
    "\n",
    "    Args:\n",
    "        model (torch.nn.Module): A PyTorch model capable of making predictions on data_loader.\n",
    "        data_loader (torch.utils.data.DataLoader): The target dataset to predict on.\n",
    "        loss_fn (torch.nn.Module): The loss function of model.\n",
    "        accuracy_fn: An accuracy function to compare the models predictions to the truth labels.\n",
    "        device (str, optional): Target device to compute on. Defaults to device.\n",
    "\n",
    "    Returns:\n",
    "        (dict): Results of model making predictions on data_loader.\n",
    "    \"\"\"\n",
    "    loss, acc = 0, 0\n",
    "    # model.to(device)  may not be necessary\n",
    "    model.eval()\n",
    "    with torch.inference_mode():\n",
    "        for X, y in data_loader:\n",
    "            # Send data to the target device\n",
    "            X, y = X.to(device), y.to(device)\n",
    "            y_pred = model(X)\n",
    "            loss += loss_fn(y_pred, y)\n",
    "            acc += accuracy_fn(y_true=y, y_pred=y_pred.argmax(dim=1))\n",
    "        \n",
    "        # Scale loss and acc\n",
    "        loss /= len(data_loader)\n",
    "        acc /= len(data_loader)\n",
    "    return {\"model_name\": model.__class__.__name__, # only works when model was created with a class\n",
    "            \"model_loss\": loss.item(),\n",
    "            \"model_acc\": acc}\n",
    "\n",
    "# Calculate model 1 results with device-agnostic code \n",
    "model_1_results = eval_model(model=model_1, data_loader=test_dataloader,\n",
    "    loss_fn=loss_fn, accuracy_fn=accuracy_fn,\n",
    "    device=device\n",
    ")\n",
    "model_1_results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "id": "a9e916cf-f873-4481-a983-bac26ce4cac2",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "a9e916cf-f873-4481-a983-bac26ce4cac2",
    "outputId": "5cdb9f7f-366c-4c14-9afa-f2d1d4e0267d"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'model_name': 'FashionMNISTModelV0',\n",
       " 'model_loss': 0.5010038018226624,\n",
       " 'model_acc': 83.12699680511182}"
      ]
     },
     "execution_count": 95,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Check baseline results\n",
    "model_0_results"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "340cbf14-e83f-4981-8a93-5fedb6b51418",
   "metadata": {
    "id": "340cbf14-e83f-4981-8a93-5fedb6b51418"
   },
   "source": [
    "Woah, in this case, it looks like adding non-linearities to our model made it perform worse than the baseline.\n",
    "\n",
    "That's a thing to note in machine learning, sometimes the thing you thought should work doesn't. \n",
    "\n",
    "And then the thing you thought might not work does.\n",
    "\n",
    "It's part science, part art.\n",
    "\n",
    "From the looks of things, it seems like our model is **overfitting** on the training data.\n",
    "\n",
    "Overfitting means our model is learning the training data well but those patterns aren't generalizing to the testing data.\n",
    "\n",
    "Two of the main to fix overfitting include:\n",
    "1. Using a smaller or different model (some models fit certain kinds of data better than others).\n",
    "2. Using a larger dataset (the more data, the more chance a model has to learn generalizable patterns).\n",
    "\n",
    "There are more, but I'm going to leave that as a challenge for you to explore.\n",
    "\n",
    "Try searching online, \"ways to prevent overfitting in machine learning\" and see what comes up.\n",
    "\n",
    "In the meantime, let's take a look at number 1: using a different model."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ac22d685-1b8d-4215-90de-c0476cb0fbdf",
   "metadata": {
    "id": "ac22d685-1b8d-4215-90de-c0476cb0fbdf"
   },
   "source": [
    "## 7. Model 2: Building a Convolutional Neural Network (CNN)\n",
    "\n",
    "Alright, time to step things up a notch.\n",
    "\n",
    "It's time to create a [Convolutional Neural Network](https://en.wikipedia.org/wiki/Convolutional_neural_network) (CNN or ConvNet).\n",
    "\n",
    "CNN's are known for their capabilities to find patterns in visual data.\n",
    "\n",
    "And since we're dealing with visual data, let's see if using a CNN model can improve upon our baseline.\n",
    "\n",
    "The CNN model we're going to be using is known as TinyVGG from the [CNN Explainer](https://poloclub.github.io/cnn-explainer/) website.\n",
    "\n",
    "It follows the typical structure of a convolutional neural network:\n",
    "\n",
    "`Input layer -> [Convolutional layer -> activation layer -> pooling layer] -> Output layer`\n",
    "\n",
    "Where the contents of `[Convolutional layer -> activation layer -> pooling layer]` can be upscaled and repeated multiple times, depending on requirements. "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9c358955-1d20-4903-b872-a239d2753d88",
   "metadata": {
    "id": "9c358955-1d20-4903-b872-a239d2753d88"
   },
   "source": [
    "### What model should I use?\n",
    "\n",
    "> **Question:** Wait, you say CNN's are good for images, are there any other model types I should be aware of?\n",
    "\n",
    "Good question.\n",
    "\n",
    "This table is a good general guide for which model to use (though there are exceptions).\n",
    "\n",
    "| **Problem type** | **Model to use (generally)** | **Code example** |\n",
    "| ----- | ----- | ----- |\n",
    "| Structured data (Excel spreadsheets, row and column data) | Gradient boosted models, Random Forests, XGBoost | [`sklearn.ensemble`](https://scikit-learn.org/stable/modules/classes.html#module-sklearn.ensemble), [XGBoost library](https://xgboost.readthedocs.io/en/stable/) |\n",
    "| Unstructured data (images, audio, language) | Convolutional Neural Networks, Transformers | [`torchvision.models`](https://pytorch.org/vision/stable/models.html), [HuggingFace Transformers](https://huggingface.co/docs/transformers/index) | \n",
    "\n",
    "> **Note:** The table above is only for reference, the model you end up using will be highly dependant on the problem you're working on and the constraints you have (amount of data, latency requirements).\n",
    "\n",
    "Enough talking about models, let's now build a CNN that replicates the model on the [CNN Explainer website](https://poloclub.github.io/cnn-explainer/).\n",
    "\n",
    "![TinyVGG architecture, as setup by CNN explainer website](https://raw.githubusercontent.com/mrdbourke/pytorch-deep-learning/main/images/03-cnn-explainer-model.png)\n",
    "\n",
    "To do so, we'll leverage the [`nn.Conv2d()`](https://pytorch.org/docs/stable/generated/torch.nn.Conv2d.html) and [`nn.MaxPool2d()`](https://pytorch.org/docs/stable/generated/torch.nn.MaxPool2d.html) layers from `torch.nn`.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "id": "dce60214-63fd-46e2-89ba-125445ac76b7",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "dce60214-63fd-46e2-89ba-125445ac76b7",
    "outputId": "5ae97191-bb41-4e58-e7f1-914b612cbb60"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "FashionMNISTModelV2(\n",
       "  (block_1): Sequential(\n",
       "    (0): Conv2d(1, 10, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))\n",
       "    (1): ReLU()\n",
       "    (2): Conv2d(10, 10, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))\n",
       "    (3): ReLU()\n",
       "    (4): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)\n",
       "  )\n",
       "  (block_2): Sequential(\n",
       "    (0): Conv2d(10, 10, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))\n",
       "    (1): ReLU()\n",
       "    (2): Conv2d(10, 10, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1))\n",
       "    (3): ReLU()\n",
       "    (4): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)\n",
       "  )\n",
       "  (classifier): Sequential(\n",
       "    (0): Flatten(start_dim=1, end_dim=-1)\n",
       "    (1): Linear(in_features=490, out_features=10, bias=True)\n",
       "  )\n",
       ")"
      ]
     },
     "execution_count": 71,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Create a convolutional neural network \n",
    "class FashionMNISTModelV2(nn.Module):\n",
    "    \"\"\"\n",
    "    Model architecture copying TinyVGG from: \n",
    "    https://poloclub.github.io/cnn-explainer/\n",
    "    \"\"\"\n",
    "    def __init__(self, input_shape: int, hidden_units: int, output_shape: int):\n",
    "        super().__init__()\n",
    "        self.block_1 = nn.Sequential(\n",
    "            nn.Conv2d(in_channels=input_shape, \n",
    "                      out_channels=hidden_units, \n",
    "                      kernel_size=3, # how big is the square that's going over the image?\n",
    "                      stride=1, # default\n",
    "                      padding=1),# options = \"valid\" (no padding) or \"same\" (output has same shape as input) or int for specific number \n",
    "            nn.ReLU(),\n",
    "            nn.Conv2d(in_channels=hidden_units, \n",
    "                      out_channels=hidden_units,\n",
    "                      kernel_size=3,\n",
    "                      stride=1,\n",
    "                      padding=1),\n",
    "            nn.ReLU(),\n",
    "            nn.MaxPool2d(kernel_size=2,\n",
    "                         stride=2) # default stride value is same as kernel_size\n",
    "        )\n",
    "        self.block_2 = nn.Sequential(\n",
    "            nn.Conv2d(hidden_units, hidden_units, 3, padding=1),\n",
    "            nn.ReLU(),\n",
    "            nn.Conv2d(hidden_units, hidden_units, 3, padding=1),\n",
    "            nn.ReLU(),\n",
    "            nn.MaxPool2d(2)\n",
    "        )\n",
    "        self.classifier = nn.Sequential(\n",
    "            nn.Flatten(),\n",
    "            # Where did this in_features shape come from? \n",
    "            # It's because each layer of our network compresses and changes the shape of our inputs data.\n",
    "            nn.Linear(in_features=hidden_units*7*7, \n",
    "                      out_features=output_shape)\n",
    "        )\n",
    "    \n",
    "    def forward(self, x: torch.Tensor):\n",
    "        x = self.block_1(x)\n",
    "        #print(x.shape) #torch.Size([1, 10, 14, 14])\n",
    "        x = self.block_2(x)\n",
    "        #print(x.shape) #torch.Size([1, 10, 7, 7])\n",
    "        x = self.classifier(x)\n",
    "        #print(x.shape)\n",
    "        return x\n",
    "\n",
    "torch.manual_seed(42)\n",
    "model_2 = FashionMNISTModelV2(input_shape=1, \n",
    "    hidden_units=10, \n",
    "    output_shape=len(class_names)).to(device)\n",
    "model_2"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0a20f25e-cc16-4f85-a69b-62008c01d0ed",
   "metadata": {
    "id": "0a20f25e-cc16-4f85-a69b-62008c01d0ed"
   },
   "source": [
    "Nice!\n",
    "\n",
    "Our biggest model yet!\n",
    "\n",
    "What we've done is a common practice in machine learning.\n",
    "\n",
    "Find a model architecture somewhere and replicate it with code. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "id": "27242622",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Total number of parameters: 7740\n"
     ]
    }
   ],
   "source": [
    "# Calculate total number of parameters in the model\n",
    "total_params = calculate_total_params(model_2)\n",
    "print(f\"Total number of parameters: {total_params}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "id": "7d3b3971",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x1de71b01ed0>"
      ]
     },
     "execution_count": 73,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "random_idx = torch.randint(0, len(train_features_batch), size=[1]).item()\n",
    "plt.imshow(train_features_batch[random_idx].squeeze(), cmap=\"gray\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "id": "4fb082a3",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "torch.Size([1, 28, 28])"
      ]
     },
     "execution_count": 74,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rand_image_tensor = torch.randn(size=(1, 28, 28))\n",
    "rand_image_tensor.shape\n",
    "#model_2(rand_image_tensor).to(device)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "id": "920a23b6",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "torch.Size([1, 28, 28])"
      ]
     },
     "execution_count": 75,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_features_batch[random_idx].shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "id": "92336d90",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tensor([[ 0.0326, -0.0636,  0.0535, -0.0278, -0.0143,  0.0251, -0.0042, -0.0150,\n",
       "          0.0306,  0.0063]], device='cuda:0', grad_fn=<AddmmBackward0>)"
      ]
     },
     "execution_count": 76,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model_2(train_features_batch[random_idx].unsqueeze(0).to(device))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "id": "a6fd3dd6",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "OrderedDict([('block_1.0.weight',\n",
       "              tensor([[[[ 0.2548,  0.2767, -0.0781],\n",
       "                        [ 0.3062, -0.0730,  0.0673],\n",
       "                        [-0.1623,  0.1958,  0.2938]]],\n",
       "              \n",
       "              \n",
       "                      [[[-0.2445,  0.2897,  0.0624],\n",
       "                        [ 0.2463,  0.0451,  0.1607],\n",
       "                        [-0.0471,  0.2570,  0.0493]]],\n",
       "              \n",
       "              \n",
       "                      [[[-0.1556,  0.0850, -0.1536],\n",
       "                        [-0.0391, -0.1354,  0.2211],\n",
       "                        [-0.2631, -0.1537, -0.0941]]],\n",
       "              \n",
       "              \n",
       "                      [[[-0.2004,  0.0315, -0.3292],\n",
       "                        [ 0.3010, -0.2832,  0.2573],\n",
       "                        [ 0.0555, -0.1082,  0.2060]]],\n",
       "              \n",
       "              \n",
       "                      [[[ 0.0520,  0.2693,  0.0364],\n",
       "                        [-0.1051,  0.0896, -0.0904],\n",
       "                        [ 0.1403,  0.2976,  0.1927]]],\n",
       "              \n",
       "              \n",
       "                      [[[-0.1457,  0.1924,  0.0596],\n",
       "                        [ 0.1693, -0.2032, -0.3300],\n",
       "                        [-0.1288, -0.2557,  0.2735]]],\n",
       "              \n",
       "              \n",
       "                      [[[ 0.0960,  0.1381,  0.1054],\n",
       "                        [-0.0058,  0.2609, -0.2368],\n",
       "                        [ 0.0210, -0.2275,  0.1028]]],\n",
       "              \n",
       "              \n",
       "                      [[[-0.1148,  0.1021, -0.0694],\n",
       "                        [ 0.2765, -0.1976, -0.1988],\n",
       "                        [-0.1988,  0.2998,  0.1111]]],\n",
       "              \n",
       "              \n",
       "                      [[[ 0.3208, -0.2751, -0.3306],\n",
       "                        [-0.2608, -0.2242,  0.1350],\n",
       "                        [ 0.1194,  0.2770, -0.1721]]],\n",
       "              \n",
       "              \n",
       "                      [[[-0.2272,  0.1769, -0.1347],\n",
       "                        [ 0.2023, -0.0791,  0.1907],\n",
       "                        [-0.2590, -0.1682,  0.1016]]]], device='cuda:0')),\n",
       "             ('block_1.0.bias',\n",
       "              tensor([ 0.0705, -0.0850,  0.1987,  0.2266, -0.2417, -0.1780,  0.3052, -0.1125,\n",
       "                      -0.1182, -0.3225], device='cuda:0')),\n",
       "             ('block_1.2.weight',\n",
       "              tensor([[[[-0.0604,  0.0263, -0.0139],\n",
       "                        [-0.0765,  0.0025, -0.0720],\n",
       "                        [-0.0894, -0.0580, -0.0923]],\n",
       "              \n",
       "                       [[-0.0671,  0.1054,  0.0199],\n",
       "                        [ 0.0325, -0.0983, -0.0692],\n",
       "                        [-0.0351,  0.0165, -0.0928]],\n",
       "              \n",
       "                       [[-0.0454, -0.0631,  0.0003],\n",
       "                        [-0.0392, -0.0073, -0.0714],\n",
       "                        [-0.0724, -0.0615, -0.0361]],\n",
       "              \n",
       "                       [[-0.0832,  0.0884, -0.0209],\n",
       "                        [ 0.0907,  0.0328, -0.0893],\n",
       "                        [ 0.0729, -0.0290, -0.0404]],\n",
       "              \n",
       "                       [[-0.0875, -0.1048,  0.0302],\n",
       "                        [-0.0230,  0.0410, -0.0865],\n",
       "                        [ 0.0783, -0.0774, -0.0182]],\n",
       "              \n",
       "                       [[ 0.0220,  0.0544,  0.0851],\n",
       "                        [ 0.0960, -0.0836,  0.0265],\n",
       "                        [-0.0453, -0.0116, -0.0789]],\n",
       "              \n",
       "                       [[ 0.0960, -0.0774,  0.0563],\n",
       "                        [ 0.0370,  0.0343, -0.0570],\n",
       "                        [ 0.0958,  0.0232,  0.0136]],\n",
       "              \n",
       "                       [[-0.0929,  0.0442, -0.0158],\n",
       "                        [-0.0483,  0.0905,  0.0235],\n",
       "                        [-0.0583, -0.0534, -0.0050]],\n",
       "              \n",
       "                       [[ 0.0589, -0.0269, -0.0601],\n",
       "                        [-0.0361, -0.0787,  0.0376],\n",
       "                        [ 0.0816, -0.0992,  0.0245]],\n",
       "              \n",
       "                       [[ 0.0545,  0.0191, -0.0375],\n",
       "                        [ 0.0550,  0.0554,  0.0394],\n",
       "                        [-0.0185, -0.0279,  0.0113]]],\n",
       "              \n",
       "              \n",
       "                      [[[-0.0186, -0.0314,  0.0674],\n",
       "                        [ 0.0906, -0.0104, -0.0236],\n",
       "                        [ 0.0015, -0.0063,  0.0253]],\n",
       "              \n",
       "                       [[ 0.0295, -0.0957, -0.0389],\n",
       "                        [ 0.0888,  0.0411, -0.0052],\n",
       "                        [-0.0636, -0.0645, -0.0944]],\n",
       "              \n",
       "                       [[-0.0344,  0.0356,  0.0672],\n",
       "                        [ 0.0487, -0.0932, -0.0634],\n",
       "                        [-0.0166,  0.1020,  0.0152]],\n",
       "              \n",
       "                       [[-0.0273,  0.0436, -0.0401],\n",
       "                        [-0.0682,  0.0769, -0.0479],\n",
       "                        [-0.0211, -0.1049,  0.0705]],\n",
       "              \n",
       "                       [[ 0.0799,  0.0384, -0.0735],\n",
       "                        [-0.1040, -0.0856,  0.0786],\n",
       "                        [ 0.0506,  0.0887,  0.0552]],\n",
       "              \n",
       "                       [[ 0.0267, -0.0010, -0.0802],\n",
       "                        [-0.0903, -0.0986,  0.0432],\n",
       "                        [-0.0518, -0.0212, -0.0607]],\n",
       "              \n",
       "                       [[-0.0192, -0.0742, -0.0689],\n",
       "                        [ 0.0350, -0.0313,  0.0651],\n",
       "                        [-0.0338, -0.0773, -0.0186]],\n",
       "              \n",
       "                       [[-0.0511, -0.0322, -0.1003],\n",
       "                        [ 0.0590, -0.0734,  0.0530],\n",
       "                        [ 0.0478,  0.0753, -0.0809]],\n",
       "              \n",
       "                       [[ 0.0758, -0.0498,  0.0391],\n",
       "                        [ 0.0990, -0.0149, -0.0008],\n",
       "                        [-0.0243, -0.0880,  0.0506]],\n",
       "              \n",
       "                       [[-0.1046,  0.0654,  0.0789],\n",
       "                        [ 0.0997, -0.0249, -0.0866],\n",
       "                        [ 0.0237,  0.0582, -0.1049]]],\n",
       "              \n",
       "              \n",
       "                      [[[-0.0239, -0.0632, -0.0092],\n",
       "                        [-0.0519, -0.0431, -0.0335],\n",
       "                        [-0.1002,  0.0865,  0.0884]],\n",
       "              \n",
       "                       [[-0.0165, -0.0120, -0.0430],\n",
       "                        [-0.0952, -0.1026,  0.0392],\n",
       "                        [-0.0579, -0.0678, -0.0082]],\n",
       "              \n",
       "                       [[-0.0351, -0.0341,  0.0034],\n",
       "                        [-0.0224, -0.0363, -0.0505],\n",
       "                        [-0.0858,  0.0884, -0.0422]],\n",
       "              \n",
       "                       [[ 0.0279, -0.0366,  0.0086],\n",
       "                        [ 0.0983,  0.0486, -0.0913],\n",
       "                        [ 0.0418,  0.1001,  0.0277]],\n",
       "              \n",
       "                       [[ 0.0707,  0.1039, -0.0162],\n",
       "                        [ 0.0219, -0.0733, -0.0217],\n",
       "                        [ 0.0781,  0.0540, -0.0667]],\n",
       "              \n",
       "                       [[-0.0845, -0.0720, -0.1040],\n",
       "                        [-0.0813, -0.0261,  0.0711],\n",
       "                        [ 0.0176, -0.0802, -0.0846]],\n",
       "              \n",
       "                       [[ 0.0524, -0.0784, -0.0130],\n",
       "                        [ 0.0506, -0.0488, -0.0115],\n",
       "                        [-0.0092, -0.0249, -0.0534]],\n",
       "              \n",
       "                       [[-0.0940, -0.0852, -0.0564],\n",
       "                        [ 0.1018, -0.0509, -0.0708],\n",
       "                        [ 0.0256,  0.0291,  0.0578]],\n",
       "              \n",
       "                       [[ 0.0801,  0.0587, -0.1045],\n",
       "                        [ 0.0093,  0.0639, -0.0097],\n",
       "                        [-0.0621,  0.1005, -0.0394]],\n",
       "              \n",
       "                       [[-0.0600, -0.0950,  0.0047],\n",
       "                        [ 0.0467,  0.0233,  0.0208],\n",
       "                        [-0.0799, -0.0984,  0.0019]]],\n",
       "              \n",
       "              \n",
       "                      [[[ 0.0961,  0.0608, -0.0614],\n",
       "                        [-0.0137, -0.0777, -0.0509],\n",
       "                        [ 0.0191,  0.0574,  0.0873]],\n",
       "              \n",
       "                       [[-0.0968,  0.0705, -0.0743],\n",
       "                        [ 0.0395,  0.0892,  0.0015],\n",
       "                        [ 0.0959, -0.0898, -0.0403]],\n",
       "              \n",
       "                       [[ 0.0615, -0.0230, -0.0216],\n",
       "                        [-0.0439,  0.0727,  0.0517],\n",
       "                        [ 0.0338, -0.0592, -0.0856]],\n",
       "              \n",
       "                       [[ 0.0114,  0.0312, -0.0487],\n",
       "                        [-0.0295,  0.0712,  0.0084],\n",
       "                        [ 0.0048, -0.0259, -0.0955]],\n",
       "              \n",
       "                       [[-0.0991, -0.0504, -0.0536],\n",
       "                        [ 0.0328, -0.0307, -0.0412],\n",
       "                        [ 0.1005,  0.0367,  0.0751]],\n",
       "              \n",
       "                       [[-0.0510, -0.0431,  0.0387],\n",
       "                        [-0.0702, -0.0689, -0.0051],\n",
       "                        [-0.0386, -0.0790,  0.0625]],\n",
       "              \n",
       "                       [[ 0.0848,  0.0171, -0.0184],\n",
       "                        [-0.0976, -0.0384,  0.0268],\n",
       "                        [ 0.0497, -0.0133, -0.0417]],\n",
       "              \n",
       "                       [[ 0.0587, -0.0839,  0.0666],\n",
       "                        [-0.0409,  0.0016, -0.0208],\n",
       "                        [ 0.0128, -0.0319,  0.0766]],\n",
       "              \n",
       "                       [[-0.0027,  0.0823,  0.1013],\n",
       "                        [-0.0514, -0.0769,  0.0846],\n",
       "                        [ 0.0826, -0.0805, -0.0081]],\n",
       "              \n",
       "                       [[-0.1039, -0.0863,  0.0204],\n",
       "                        [ 0.0280,  0.0223, -0.0287],\n",
       "                        [ 0.0972,  0.0151, -0.0622]]],\n",
       "              \n",
       "              \n",
       "                      [[[-0.0060,  0.0253,  0.0369],\n",
       "                        [-0.0745,  0.0395, -0.0539],\n",
       "                        [-0.0876, -0.0576,  0.1017]],\n",
       "              \n",
       "                       [[ 0.0901,  0.0944,  0.0619],\n",
       "                        [ 0.0796, -0.0141, -0.0580],\n",
       "                        [ 0.0527, -0.0546, -0.0711]],\n",
       "              \n",
       "                       [[-0.0337,  0.0221,  0.0543],\n",
       "                        [-0.0409, -0.0620,  0.0142],\n",
       "                        [-0.0621, -0.0686,  0.0549]],\n",
       "              \n",
       "                       [[-0.0177,  0.0963,  0.1025],\n",
       "                        [ 0.0315,  0.0363,  0.0243],\n",
       "                        [ 0.0017, -0.0077,  0.0014]],\n",
       "              \n",
       "                       [[ 0.0394,  0.0980, -0.0273],\n",
       "                        [-0.0446, -0.0255, -0.0509],\n",
       "                        [ 0.0179,  0.0787,  0.0824]],\n",
       "              \n",
       "                       [[ 0.0484, -0.0776, -0.0566],\n",
       "                        [-0.0232, -0.0194,  0.0087],\n",
       "                        [-0.0968,  0.0328, -0.0804]],\n",
       "              \n",
       "                       [[-0.0667, -0.0876,  0.0918],\n",
       "                        [-0.0998,  0.0795, -0.0035],\n",
       "                        [-0.0123,  0.0659, -0.0097]],\n",
       "              \n",
       "                       [[ 0.0661,  0.0762, -0.0915],\n",
       "                        [ 0.0406,  0.0199,  0.0227],\n",
       "                        [ 0.0154,  0.0288, -0.0507]],\n",
       "              \n",
       "                       [[-0.0135,  0.1002,  0.0708],\n",
       "                        [-0.0040, -0.0991,  0.0046],\n",
       "                        [-0.0718,  0.0857, -0.0640]],\n",
       "              \n",
       "                       [[-0.0076, -0.0234,  0.0188],\n",
       "                        [ 0.0992,  0.0100,  0.0610],\n",
       "                        [ 0.0818,  0.0851, -0.0364]]],\n",
       "              \n",
       "              \n",
       "                      [[[-0.0236,  0.0508, -0.0288],\n",
       "                        [ 0.0494, -0.0230, -0.0715],\n",
       "                        [ 0.0429,  0.0162,  0.0470]],\n",
       "              \n",
       "                       [[ 0.1047,  0.0720,  0.0999],\n",
       "                        [ 0.0056, -0.0907, -0.0739],\n",
       "                        [-0.0655, -0.0929, -0.0528]],\n",
       "              \n",
       "                       [[-0.0970, -0.0973, -0.0630],\n",
       "                        [-0.1039, -0.0647,  0.0402],\n",
       "                        [ 0.0879, -0.0314, -0.0307]],\n",
       "              \n",
       "                       [[ 0.0563, -0.0520, -0.0498],\n",
       "                        [ 0.0649, -0.0918,  0.0129],\n",
       "                        [ 0.0931,  0.0181,  0.0287]],\n",
       "              \n",
       "                       [[-0.0614, -0.0015,  0.0058],\n",
       "                        [ 0.0259,  0.0410,  0.0916],\n",
       "                        [-0.0805,  0.0032, -0.0527]],\n",
       "              \n",
       "                       [[-0.0834, -0.0084, -0.0928],\n",
       "                        [ 0.0736,  0.0122, -0.0568],\n",
       "                        [ 0.0551, -0.0998, -0.0408]],\n",
       "              \n",
       "                       [[-0.0205, -0.0896, -0.0670],\n",
       "                        [-0.0172,  0.0800,  0.1018],\n",
       "                        [ 0.0671, -0.0629, -0.0690]],\n",
       "              \n",
       "                       [[ 0.0920,  0.0373,  0.0028],\n",
       "                        [ 0.0143, -0.0847, -0.0352],\n",
       "                        [ 0.1015, -0.0260, -0.0053]],\n",
       "              \n",
       "                       [[-0.0875, -0.0590, -0.0022],\n",
       "                        [-0.0655, -0.0131,  0.0429],\n",
       "                        [-0.1031,  0.0313, -0.0697]],\n",
       "              \n",
       "                       [[-0.0514,  0.0405,  0.0838],\n",
       "                        [-0.0288, -0.0433, -0.0953],\n",
       "                        [-0.0544, -0.0923, -0.0241]]],\n",
       "              \n",
       "              \n",
       "                      [[[ 0.0215, -0.0988,  0.0920],\n",
       "                        [ 0.0661, -0.1032, -0.0503],\n",
       "                        [ 0.0344, -0.0217, -0.0115]],\n",
       "              \n",
       "                       [[-0.0476,  0.0847, -0.0589],\n",
       "                        [ 0.0874,  0.0068,  0.0212],\n",
       "                        [ 0.0822, -0.0174, -0.0600]],\n",
       "              \n",
       "                       [[-0.0170,  0.0855, -0.0782],\n",
       "                        [ 0.0239, -0.1036,  0.0553],\n",
       "                        [ 0.0389,  0.0045,  0.0452]],\n",
       "              \n",
       "                       [[ 0.0001,  0.0583, -0.0834],\n",
       "                        [-0.0155,  0.0468,  0.1050],\n",
       "                        [ 0.0537, -0.0767,  0.0811]],\n",
       "              \n",
       "                       [[-0.0235, -0.0225, -0.0958],\n",
       "                        [-0.0166,  0.0746,  0.0147],\n",
       "                        [-0.0614,  0.0324, -0.0338]],\n",
       "              \n",
       "                       [[ 0.0962, -0.0915, -0.0333],\n",
       "                        [-0.1018, -0.0415,  0.0332],\n",
       "                        [ 0.1015,  0.0177,  0.1033]],\n",
       "              \n",
       "                       [[ 0.0206,  0.0609,  0.0845],\n",
       "                        [ 0.0881, -0.0590,  0.0969],\n",
       "                        [ 0.0639, -0.0493, -0.0503]],\n",
       "              \n",
       "                       [[-0.0884,  0.0265, -0.0854],\n",
       "                        [ 0.0445,  0.0333, -0.0916],\n",
       "                        [ 0.0287, -0.0086,  0.0482]],\n",
       "              \n",
       "                       [[ 0.0605, -0.1048,  0.0967],\n",
       "                        [ 0.0884,  0.0419, -0.0963],\n",
       "                        [-0.0377, -0.0305, -0.0271]],\n",
       "              \n",
       "                       [[ 0.0594,  0.0383,  0.0835],\n",
       "                        [-0.0395,  0.0355,  0.0375],\n",
       "                        [-0.0878, -0.1022, -0.0547]]],\n",
       "              \n",
       "              \n",
       "                      [[[ 0.0722, -0.0992, -0.0918],\n",
       "                        [ 0.0591,  0.0569,  0.0867],\n",
       "                        [-0.0796, -0.0771,  0.0541]],\n",
       "              \n",
       "                       [[ 0.0917,  0.0631,  0.0165],\n",
       "                        [ 0.0347,  0.1000, -0.0680],\n",
       "                        [-0.0479,  0.0737, -0.0721]],\n",
       "              \n",
       "                       [[-0.0581,  0.0769,  0.0333],\n",
       "                        [ 0.0341, -0.0447, -0.0015],\n",
       "                        [ 0.0965, -0.0633,  0.0008]],\n",
       "              \n",
       "                       [[ 0.0501, -0.0728,  0.1024],\n",
       "                        [-0.0527, -0.0253, -0.0285],\n",
       "                        [-0.0687, -0.1034,  0.0594]],\n",
       "              \n",
       "                       [[ 0.0280, -0.0987, -0.0678],\n",
       "                        [ 0.1042,  0.0403,  0.0423],\n",
       "                        [-0.0631, -0.0462, -0.0159]],\n",
       "              \n",
       "                       [[-0.0193, -0.0722,  0.0087],\n",
       "                        [ 0.0105, -0.0133,  0.0146],\n",
       "                        [-0.0418,  0.0274,  0.0398]],\n",
       "              \n",
       "                       [[-0.0555, -0.1045,  0.0552],\n",
       "                        [ 0.0251, -0.0536,  0.1016],\n",
       "                        [-0.0477,  0.0712,  0.0535]],\n",
       "              \n",
       "                       [[-0.0884,  0.0680, -0.0969],\n",
       "                        [-0.0584, -0.0176, -0.0711],\n",
       "                        [ 0.1030, -0.0211,  0.0419]],\n",
       "              \n",
       "                       [[-0.0941,  0.0607, -0.0328],\n",
       "                        [-0.0802,  0.0154,  0.0511],\n",
       "                        [ 0.0912, -0.0644, -0.0519]],\n",
       "              \n",
       "                       [[ 0.0203,  0.0286,  0.0405],\n",
       "                        [ 0.0579, -0.0239,  0.0586],\n",
       "                        [ 0.0777, -0.0275,  0.0750]]],\n",
       "              \n",
       "              \n",
       "                      [[[ 0.0515,  0.0930, -0.0599],\n",
       "                        [-0.0521, -0.0305,  0.0053],\n",
       "                        [ 0.0633, -0.0602,  0.0528]],\n",
       "              \n",
       "                       [[-0.0378,  0.0637, -0.0050],\n",
       "                        [-0.0923, -0.0580, -0.0763],\n",
       "                        [ 0.0523, -0.0707, -0.0088]],\n",
       "              \n",
       "                       [[ 0.0227, -0.0578,  0.0304],\n",
       "                        [-0.1029, -0.0754, -0.0955],\n",
       "                        [-0.0319, -0.0384,  0.0151]],\n",
       "              \n",
       "                       [[-0.0195,  0.0496,  0.0966],\n",
       "                        [ 0.0378, -0.0415, -0.0987],\n",
       "                        [ 0.0382, -0.0522,  0.0536]],\n",
       "              \n",
       "                       [[ 0.0705,  0.0407,  0.0989],\n",
       "                        [ 0.1001,  0.0223, -0.0768],\n",
       "                        [ 0.0942, -0.0500, -0.0498]],\n",
       "              \n",
       "                       [[ 0.0882,  0.0817,  0.0318],\n",
       "                        [ 0.0066, -0.0887, -0.0109],\n",
       "                        [ 0.1011,  0.0268,  0.0090]],\n",
       "              \n",
       "                       [[-0.0219, -0.0368,  0.0628],\n",
       "                        [ 0.0065,  0.0686, -0.0187],\n",
       "                        [ 0.0461,  0.0435,  0.0168]],\n",
       "              \n",
       "                       [[ 0.0662,  0.0661,  0.0977],\n",
       "                        [ 0.0810, -0.0270, -0.0892],\n",
       "                        [ 0.0193, -0.0009, -0.0275]],\n",
       "              \n",
       "                       [[-0.0177,  0.0050,  0.0769],\n",
       "                        [ 0.0329, -0.0374, -0.0433],\n",
       "                        [-0.0261, -0.0407,  0.0948]],\n",
       "              \n",
       "                       [[ 0.0558,  0.0952,  0.0003],\n",
       "                        [ 0.0213,  0.0366, -0.0998],\n",
       "                        [ 0.0094, -0.0071, -0.0591]]],\n",
       "              \n",
       "              \n",
       "                      [[[-0.0818,  0.0933,  0.0857],\n",
       "                        [ 0.0489,  0.1006, -0.0428],\n",
       "                        [-0.0182,  0.0399, -0.0174]],\n",
       "              \n",
       "                       [[-0.0207, -0.0871,  0.0283],\n",
       "                        [-0.0637,  0.0038,  0.1028],\n",
       "                        [-0.0324, -0.0332,  0.0636]],\n",
       "              \n",
       "                       [[-0.0388, -0.0091,  0.0984],\n",
       "                        [-0.0432, -0.0754, -0.0590],\n",
       "                        [-0.0292, -0.0500, -0.0547]],\n",
       "              \n",
       "                       [[ 0.0426,  0.0179, -0.0337],\n",
       "                        [-0.0819, -0.0332, -0.0445],\n",
       "                        [-0.0343, -0.0951,  0.0227]],\n",
       "              \n",
       "                       [[-0.0774, -0.0821, -0.0861],\n",
       "                        [ 0.0440, -0.0635, -0.0435],\n",
       "                        [ 0.0826,  0.0560,  0.0604]],\n",
       "              \n",
       "                       [[-0.1001, -0.0756, -0.0398],\n",
       "                        [ 0.0871,  0.0108, -0.0788],\n",
       "                        [ 0.0007, -0.0819, -0.0231]],\n",
       "              \n",
       "                       [[-0.0290,  0.0912,  0.0326],\n",
       "                        [-0.0184,  0.0178, -0.0304],\n",
       "                        [ 0.0414,  0.0417,  0.0283]],\n",
       "              \n",
       "                       [[-0.0411,  0.0899, -0.0152],\n",
       "                        [-0.0410,  0.0660,  0.0859],\n",
       "                        [ 0.1049,  0.0312, -0.0359]],\n",
       "              \n",
       "                       [[ 0.0535,  0.0904, -0.1034],\n",
       "                        [-0.0131, -0.0719,  0.0196],\n",
       "                        [ 0.0436, -0.0218, -0.0088]],\n",
       "              \n",
       "                       [[ 0.0474, -0.0177, -0.0885],\n",
       "                        [ 0.0843, -0.0531, -0.0116],\n",
       "                        [ 0.0099, -0.0063, -0.0992]]]], device='cuda:0')),\n",
       "             ('block_1.2.bias',\n",
       "              tensor([ 0.0484, -0.0479, -0.0547,  0.0252, -0.0550, -0.0487, -0.0355, -0.0396,\n",
       "                      -0.0440, -0.0284], device='cuda:0')),\n",
       "             ('block_2.0.weight',\n",
       "              tensor([[[[ 2.7393e-02, -8.5299e-02, -6.3802e-02],\n",
       "                        [ 1.5381e-03,  1.4659e-02,  5.8217e-02],\n",
       "                        [-7.4044e-02,  3.3646e-02,  5.9914e-02]],\n",
       "              \n",
       "                       [[ 5.8530e-02, -9.8180e-02, -4.0225e-02],\n",
       "                        [-9.0606e-02, -6.6704e-02,  5.8711e-02],\n",
       "                        [-1.5740e-02,  4.4769e-02, -6.1876e-02]],\n",
       "              \n",
       "                       [[ 1.6018e-02, -6.3758e-02,  5.2693e-02],\n",
       "                        [-4.6104e-02, -2.6432e-02, -9.1456e-02],\n",
       "                        [ 3.4823e-04,  1.0008e-01,  5.1163e-02]],\n",
       "              \n",
       "                       [[-5.6240e-02,  1.4176e-03, -1.1558e-02],\n",
       "                        [-8.4862e-02,  8.2650e-02,  1.6993e-03],\n",
       "                        [ 2.2199e-02, -4.2567e-02, -4.9323e-02]],\n",
       "              \n",
       "                       [[ 1.7381e-02,  3.8971e-02,  2.3643e-02],\n",
       "                        [-5.0801e-02,  1.0234e-01, -1.5517e-02],\n",
       "                        [-6.4554e-02, -4.9301e-02,  1.0377e-01]],\n",
       "              \n",
       "                       [[ 5.0738e-06, -1.4309e-02, -4.3867e-02],\n",
       "                        [-2.7633e-02, -8.8779e-02, -8.3767e-02],\n",
       "                        [ 6.1695e-02,  9.0172e-02,  1.0059e-01]],\n",
       "              \n",
       "                       [[-7.6099e-02,  5.7012e-02, -6.5245e-02],\n",
       "                        [ 6.2883e-02,  7.6058e-02,  8.1573e-02],\n",
       "                        [ 7.5900e-02,  6.5941e-02,  2.0517e-03]],\n",
       "              \n",
       "                       [[ 4.8434e-02, -3.7712e-02,  4.5899e-02],\n",
       "                        [-3.3879e-02, -1.7700e-03, -9.1746e-02],\n",
       "                        [-2.7562e-02, -5.5432e-02, -3.5557e-02]],\n",
       "              \n",
       "                       [[-6.7313e-02, -9.4810e-02,  6.8639e-03],\n",
       "                        [ 6.8408e-02,  9.6001e-02,  6.1512e-02],\n",
       "                        [-5.4638e-02, -1.0425e-01,  3.9983e-02]],\n",
       "              \n",
       "                       [[ 5.9062e-02, -9.0495e-02,  3.7798e-02],\n",
       "                        [ 8.9121e-02,  6.3853e-03, -6.3505e-02],\n",
       "                        [ 8.6423e-02,  4.5011e-02,  6.9802e-02]]],\n",
       "              \n",
       "              \n",
       "                      [[[-7.1287e-02,  6.1342e-02, -7.2002e-02],\n",
       "                        [ 1.0430e-01, -4.4662e-02,  6.3516e-02],\n",
       "                        [ 2.1107e-02,  2.7935e-02, -1.6165e-02]],\n",
       "              \n",
       "                       [[ 4.3295e-02, -4.3932e-02, -9.9357e-02],\n",
       "                        [-4.0499e-02,  8.2592e-02, -2.7751e-02],\n",
       "                        [ 3.3132e-02, -3.8973e-02,  7.9073e-02]],\n",
       "              \n",
       "                       [[ 6.3086e-02,  3.7211e-02, -5.3881e-02],\n",
       "                        [-8.6133e-02,  3.9686e-03, -6.1839e-02],\n",
       "                        [ 8.6667e-02, -1.0130e-01,  4.7104e-02]],\n",
       "              \n",
       "                       [[ 1.0508e-01,  5.2792e-02,  3.5942e-02],\n",
       "                        [-1.0142e-01,  1.0139e-01, -1.8030e-02],\n",
       "                        [-9.8495e-02,  1.0406e-01, -4.2894e-02]],\n",
       "              \n",
       "                       [[-7.4575e-03,  9.6479e-02, -7.3070e-02],\n",
       "                        [-7.4576e-02,  1.7141e-02, -1.4109e-02],\n",
       "                        [ 2.4280e-02, -8.8407e-02,  3.1524e-03]],\n",
       "              \n",
       "                       [[-4.6882e-02, -5.1820e-02, -9.6517e-02],\n",
       "                        [ 5.5890e-02,  2.0306e-02, -8.9118e-02],\n",
       "                        [ 8.3648e-02,  3.1794e-02,  1.9560e-02]],\n",
       "              \n",
       "                       [[-6.1890e-02,  1.5896e-02,  1.0157e-01],\n",
       "                        [ 7.2299e-02, -8.2100e-02,  9.6220e-02],\n",
       "                        [ 8.1702e-03,  5.0698e-02,  8.1869e-02]],\n",
       "              \n",
       "                       [[ 8.9862e-02, -8.2170e-02,  9.2303e-02],\n",
       "                        [-7.1591e-02,  7.9021e-03, -7.3656e-02],\n",
       "                        [-2.3109e-02, -4.7901e-03, -1.2611e-02]],\n",
       "              \n",
       "                       [[-1.6652e-02,  8.3137e-03,  1.0398e-01],\n",
       "                        [ 6.1244e-02,  5.8973e-02,  4.2190e-02],\n",
       "                        [ 8.1606e-02, -4.8645e-03,  8.3813e-03]],\n",
       "              \n",
       "                       [[ 2.1693e-02, -9.1931e-02, -8.4913e-02],\n",
       "                        [ 1.2923e-02, -4.1241e-02, -1.9342e-03],\n",
       "                        [-2.4187e-02,  1.6408e-02,  6.8581e-02]]],\n",
       "              \n",
       "              \n",
       "                      [[[-3.4958e-02,  8.4418e-02,  8.3227e-02],\n",
       "                        [-8.0901e-02, -8.1400e-02, -8.5284e-02],\n",
       "                        [-5.7766e-02, -4.1033e-02, -7.9341e-03]],\n",
       "              \n",
       "                       [[-2.5635e-02, -5.3258e-02, -3.3488e-02],\n",
       "                        [-3.8131e-02,  1.0341e-01, -3.9068e-02],\n",
       "                        [-7.5473e-02,  4.3818e-02, -6.0886e-03]],\n",
       "              \n",
       "                       [[ 8.0698e-02,  6.5863e-02,  9.6843e-02],\n",
       "                        [-7.7197e-02,  6.7764e-02,  8.8464e-02],\n",
       "                        [-5.2054e-02,  9.6890e-02,  7.9019e-02]],\n",
       "              \n",
       "                       [[ 1.1544e-03,  5.0823e-02, -3.6853e-02],\n",
       "                        [-9.1936e-02,  2.6645e-02,  3.1425e-02],\n",
       "                        [-6.8891e-02,  5.1123e-02, -9.0043e-02]],\n",
       "              \n",
       "                       [[ 9.0718e-02,  1.0208e-01,  2.8699e-02],\n",
       "                        [-6.6137e-02,  5.1300e-02,  1.7963e-02],\n",
       "                        [ 2.8663e-02,  3.4643e-02,  8.0254e-02]],\n",
       "              \n",
       "                       [[-4.5309e-02, -2.3711e-02,  2.8746e-02],\n",
       "                        [ 1.1486e-02,  8.5000e-02, -5.5365e-02],\n",
       "                        [-3.8387e-03,  1.9696e-02, -2.7996e-02]],\n",
       "              \n",
       "                       [[ 7.1859e-02,  1.1530e-02, -9.7422e-02],\n",
       "                        [-1.1420e-02, -4.7809e-02,  1.0243e-02],\n",
       "                        [-1.2250e-02, -1.0456e-01, -1.9208e-02]],\n",
       "              \n",
       "                       [[-1.0096e-02, -3.1083e-02,  9.6848e-02],\n",
       "                        [-2.3000e-02,  6.7717e-02,  2.6112e-02],\n",
       "                        [-8.8979e-02,  2.4770e-02,  8.7356e-02]],\n",
       "              \n",
       "                       [[-6.8948e-02, -6.8134e-02,  1.0318e-01],\n",
       "                        [ 8.4697e-02, -5.8807e-02,  6.3429e-02],\n",
       "                        [-1.3485e-02, -1.0393e-01,  7.9198e-03]],\n",
       "              \n",
       "                       [[ 3.4057e-02, -3.1619e-02,  3.6670e-02],\n",
       "                        [-9.0136e-02,  7.3050e-02,  8.9865e-02],\n",
       "                        [ 5.8130e-02,  1.7866e-02,  3.4716e-02]]],\n",
       "              \n",
       "              \n",
       "                      [[[-7.6269e-02, -2.6339e-02, -1.0063e-02],\n",
       "                        [-5.8659e-02, -7.7857e-02,  7.0900e-02],\n",
       "                        [ 7.1535e-02, -9.5731e-02,  3.3542e-02]],\n",
       "              \n",
       "                       [[ 4.2881e-02,  1.0014e-01,  6.0985e-02],\n",
       "                        [ 9.6907e-02, -3.4510e-02,  7.3827e-02],\n",
       "                        [ 8.5740e-02, -9.9541e-02, -8.4613e-02]],\n",
       "              \n",
       "                       [[ 2.1335e-02,  5.7557e-02, -5.2369e-02],\n",
       "                        [ 1.1609e-02, -1.5303e-04,  2.6680e-02],\n",
       "                        [-5.6642e-02,  5.9455e-02,  7.0098e-02]],\n",
       "              \n",
       "                       [[-7.3139e-02,  1.0211e-03,  2.9247e-04],\n",
       "                        [ 3.3849e-02,  9.8198e-02,  3.0913e-02],\n",
       "                        [-2.3951e-02,  9.4672e-02, -4.0112e-02]],\n",
       "              \n",
       "                       [[-3.0608e-02,  7.1969e-03, -8.0270e-02],\n",
       "                        [ 1.1470e-02, -7.1518e-02,  1.0838e-02],\n",
       "                        [ 1.0099e-02,  1.4591e-02, -8.8891e-02]],\n",
       "              \n",
       "                       [[-1.0012e-01,  4.8501e-02,  9.0399e-02],\n",
       "                        [-9.3537e-02,  3.9043e-02, -7.7594e-02],\n",
       "                        [ 6.6082e-03,  9.8068e-02,  7.9965e-02]],\n",
       "              \n",
       "                       [[-7.7069e-02,  6.5203e-02,  5.5057e-02],\n",
       "                        [-1.6169e-04,  1.0211e-01, -4.1866e-02],\n",
       "                        [-2.4530e-02, -5.3275e-02,  1.5168e-02]],\n",
       "              \n",
       "                       [[ 2.7911e-02,  8.3990e-03, -5.9307e-02],\n",
       "                        [-4.7452e-02,  3.5855e-02, -9.2426e-02],\n",
       "                        [-1.6416e-02, -2.3350e-03, -4.2708e-02]],\n",
       "              \n",
       "                       [[ 3.8360e-02,  6.7940e-03,  7.4004e-02],\n",
       "                        [-9.3616e-03, -6.6528e-02,  7.4477e-02],\n",
       "                        [ 1.4720e-02, -3.0189e-02, -6.9476e-02]],\n",
       "              \n",
       "                       [[ 2.4707e-02, -1.0053e-01,  2.7762e-02],\n",
       "                        [ 5.2119e-02, -9.2465e-02, -6.9009e-02],\n",
       "                        [-7.5781e-02,  8.8597e-02,  8.9611e-02]]],\n",
       "              \n",
       "              \n",
       "                      [[[ 6.5987e-03,  9.8959e-02, -3.5239e-02],\n",
       "                        [-1.0233e-01,  3.6819e-02,  3.7343e-02],\n",
       "                        [ 1.0334e-01, -3.0510e-05,  8.0785e-02]],\n",
       "              \n",
       "                       [[ 6.4612e-02,  7.6292e-02, -1.0460e-01],\n",
       "                        [ 8.6800e-02, -8.9856e-02,  9.4501e-02],\n",
       "                        [-4.3682e-03, -9.3415e-02,  2.9314e-02]],\n",
       "              \n",
       "                       [[-2.1456e-02, -9.4678e-02, -3.8215e-02],\n",
       "                        [ 1.0868e-02,  8.2098e-02, -3.2406e-02],\n",
       "                        [ 6.2610e-02,  1.3200e-02,  3.5531e-03]],\n",
       "              \n",
       "                       [[ 2.0170e-02, -6.9177e-02, -8.7616e-02],\n",
       "                        [-3.3121e-02, -9.8226e-02, -4.9158e-02],\n",
       "                        [ 4.8494e-03, -6.9424e-02, -4.3723e-02]],\n",
       "              \n",
       "                       [[-1.8941e-02, -1.2144e-02, -5.8187e-02],\n",
       "                        [ 5.0650e-03, -1.4795e-02,  3.0147e-02],\n",
       "                        [ 4.7611e-03, -5.2638e-02, -3.6291e-02]],\n",
       "              \n",
       "                       [[-1.2149e-03, -6.5774e-02,  8.2520e-03],\n",
       "                        [-7.4425e-03,  4.0897e-02,  2.4947e-02],\n",
       "                        [ 7.8887e-02, -3.4749e-03, -7.7887e-02]],\n",
       "              \n",
       "                       [[ 4.7119e-02, -7.1240e-02, -1.4489e-02],\n",
       "                        [-3.4132e-02, -3.9997e-02, -3.9000e-02],\n",
       "                        [ 9.6863e-02,  6.0342e-02,  2.9213e-02]],\n",
       "              \n",
       "                       [[ 9.8975e-02, -9.5524e-02,  1.7010e-02],\n",
       "                        [ 6.7481e-02,  7.0022e-02, -8.3890e-02],\n",
       "                        [ 3.7514e-02, -6.0050e-02, -4.1187e-03]],\n",
       "              \n",
       "                       [[-2.1996e-02, -8.8013e-02, -1.0055e-01],\n",
       "                        [-6.9349e-02,  4.7832e-02,  4.8218e-02],\n",
       "                        [-9.1681e-02, -3.9586e-02,  1.7218e-03]],\n",
       "              \n",
       "                       [[-9.1135e-02,  5.9393e-02,  9.5473e-02],\n",
       "                        [ 1.8643e-02, -7.8321e-02,  2.4580e-02],\n",
       "                        [ 3.8265e-02,  8.3468e-02, -5.6085e-02]]],\n",
       "              \n",
       "              \n",
       "                      [[[-9.4437e-02,  4.6312e-02,  6.5624e-03],\n",
       "                        [-3.4345e-02, -4.4169e-02, -5.4351e-02],\n",
       "                        [ 8.5328e-02, -1.8187e-02,  7.6022e-02]],\n",
       "              \n",
       "                       [[ 9.4094e-02,  1.3353e-02,  2.2454e-02],\n",
       "                        [-7.1789e-03,  7.2397e-02, -9.4983e-02],\n",
       "                        [ 4.1919e-02, -1.7174e-02,  4.8132e-02]],\n",
       "              \n",
       "                       [[-4.6949e-04, -3.9029e-02, -1.1379e-02],\n",
       "                        [ 5.6920e-02, -7.3210e-02, -6.6629e-02],\n",
       "                        [-2.3611e-02, -3.8235e-02,  4.1409e-02]],\n",
       "              \n",
       "                       [[ 7.0937e-02, -1.1289e-02,  9.9672e-02],\n",
       "                        [-4.4042e-02, -5.9151e-02, -4.7191e-02],\n",
       "                        [-7.2624e-02, -7.3885e-02, -9.3921e-02]],\n",
       "              \n",
       "                       [[-9.3422e-02,  2.7512e-02,  6.4284e-02],\n",
       "                        [ 9.8963e-02,  8.9787e-02, -6.0709e-03],\n",
       "                        [ 2.0454e-02, -6.3068e-02,  4.0743e-02]],\n",
       "              \n",
       "                       [[-1.0107e-01,  4.9719e-02,  1.9334e-02],\n",
       "                        [ 3.2393e-02,  3.8595e-02, -4.8394e-02],\n",
       "                        [ 9.0452e-02,  5.0307e-02,  6.9243e-02]],\n",
       "              \n",
       "                       [[ 1.3922e-02,  6.6196e-02,  7.0941e-02],\n",
       "                        [ 4.7775e-02,  8.0297e-02, -1.9119e-02],\n",
       "                        [ 6.9310e-02,  2.4286e-02,  6.3424e-02]],\n",
       "              \n",
       "                       [[ 1.0267e-01,  2.3869e-02, -3.9124e-02],\n",
       "                        [-1.0488e-02,  2.9676e-02,  1.7773e-02],\n",
       "                        [-2.8795e-02,  8.2590e-02,  6.3331e-02]],\n",
       "              \n",
       "                       [[-6.5475e-02, -8.5889e-03, -1.0119e-02],\n",
       "                        [-6.6063e-02,  1.5374e-02, -3.2360e-02],\n",
       "                        [-5.4419e-02, -3.3894e-02, -3.7584e-02]],\n",
       "              \n",
       "                       [[ 1.0084e-01,  4.0432e-02,  1.0373e-01],\n",
       "                        [ 2.8903e-02,  2.3868e-02,  4.3333e-02],\n",
       "                        [ 1.8092e-02, -8.2722e-02, -6.2334e-02]]],\n",
       "              \n",
       "              \n",
       "                      [[[-2.5538e-02,  1.5846e-03,  3.9709e-02],\n",
       "                        [ 4.0588e-02,  8.3623e-02,  2.1458e-02],\n",
       "                        [-3.5975e-02, -7.9271e-02, -7.7203e-02]],\n",
       "              \n",
       "                       [[-6.2965e-02,  3.1792e-02,  5.6950e-02],\n",
       "                        [ 9.2224e-02, -3.3342e-02, -8.3150e-03],\n",
       "                        [-3.1303e-02, -3.8517e-04,  3.3837e-02]],\n",
       "              \n",
       "                       [[-2.3160e-03,  4.8799e-03,  1.3354e-02],\n",
       "                        [ 3.9256e-02, -3.1981e-02, -6.2855e-02],\n",
       "                        [ 2.4869e-02, -1.2481e-02, -4.7753e-02]],\n",
       "              \n",
       "                       [[ 4.4268e-02,  9.5597e-04, -1.5333e-02],\n",
       "                        [-5.1027e-02, -1.3868e-02, -8.9632e-02],\n",
       "                        [ 2.3980e-02,  1.5818e-03,  6.3966e-02]],\n",
       "              \n",
       "                       [[ 6.8063e-03,  8.4277e-03,  2.8715e-02],\n",
       "                        [ 8.0210e-02, -4.9812e-02,  6.2930e-02],\n",
       "                        [ 2.5779e-02, -7.0320e-02,  3.6702e-02]],\n",
       "              \n",
       "                       [[-6.3217e-02, -3.3181e-02, -5.0245e-02],\n",
       "                        [-7.1711e-02,  8.3017e-02, -9.4217e-02],\n",
       "                        [ 5.2706e-02, -9.4870e-02, -1.2829e-02]],\n",
       "              \n",
       "                       [[ 6.2868e-03,  7.4937e-02, -3.8147e-02],\n",
       "                        [ 3.0340e-02,  1.6329e-02,  6.2021e-02],\n",
       "                        [ 6.2668e-03,  3.9470e-02, -6.3677e-02]],\n",
       "              \n",
       "                       [[-7.3250e-02,  9.3928e-02, -7.6808e-02],\n",
       "                        [-1.7945e-02, -1.2742e-02,  1.0308e-01],\n",
       "                        [-2.2780e-02, -8.0249e-02, -2.6721e-02]],\n",
       "              \n",
       "                       [[ 5.4372e-02,  4.1773e-02,  8.7204e-02],\n",
       "                        [-2.1579e-02,  4.9653e-02, -9.9194e-02],\n",
       "                        [ 4.0787e-02,  4.8432e-02,  6.7998e-02]],\n",
       "              \n",
       "                       [[-6.0446e-02, -2.8142e-02,  2.5502e-02],\n",
       "                        [-7.4905e-02, -8.3851e-02, -1.0141e-01],\n",
       "                        [ 5.8842e-03,  6.5458e-02,  2.7075e-02]]],\n",
       "              \n",
       "              \n",
       "                      [[[ 6.4263e-03,  3.6727e-02, -6.6240e-02],\n",
       "                        [ 1.1113e-02, -2.6186e-02, -5.2193e-02],\n",
       "                        [ 9.0902e-02, -8.1550e-02,  1.5448e-02]],\n",
       "              \n",
       "                       [[-9.2624e-02, -3.5762e-03, -4.6840e-02],\n",
       "                        [ 3.4695e-02, -5.9191e-02,  6.7466e-02],\n",
       "                        [-8.5536e-02,  6.3313e-02, -7.9181e-02]],\n",
       "              \n",
       "                       [[ 5.6456e-02, -4.4384e-02, -2.4556e-04],\n",
       "                        [-1.9238e-02,  6.8414e-02,  3.4546e-02],\n",
       "                        [-9.2887e-02,  9.6914e-03, -7.2718e-02]],\n",
       "              \n",
       "                       [[ 7.8800e-02,  1.7319e-02, -2.7109e-02],\n",
       "                        [-5.3777e-02,  3.6485e-02, -6.3129e-02],\n",
       "                        [ 4.9992e-02,  5.7519e-02,  6.4701e-02]],\n",
       "              \n",
       "                       [[ 2.7537e-02, -9.2272e-02,  7.5823e-02],\n",
       "                        [-3.2700e-02, -3.1163e-02, -1.1325e-02],\n",
       "                        [ 7.7068e-02,  8.1052e-02,  1.6276e-02]],\n",
       "              \n",
       "                       [[ 5.0296e-02, -9.8241e-02,  2.4901e-04],\n",
       "                        [-9.3254e-02,  3.5876e-02, -7.5099e-02],\n",
       "                        [-3.7568e-02,  7.3684e-02,  1.0074e-01]],\n",
       "              \n",
       "                       [[-6.3286e-02, -5.8503e-02,  1.3055e-02],\n",
       "                        [ 4.1437e-02, -1.7168e-02, -3.2918e-02],\n",
       "                        [-6.9237e-02,  4.4997e-02,  1.0328e-01]],\n",
       "              \n",
       "                       [[-5.1026e-02,  4.9718e-02,  5.1481e-02],\n",
       "                        [ 8.4728e-02, -1.2001e-02,  3.3202e-03],\n",
       "                        [ 7.7444e-02,  6.6631e-02,  1.0411e-01]],\n",
       "              \n",
       "                       [[-3.0207e-02,  4.1709e-02,  7.3605e-02],\n",
       "                        [-7.1553e-02,  2.0940e-02, -2.3586e-02],\n",
       "                        [ 6.7760e-02, -4.7342e-02,  7.3933e-03]],\n",
       "              \n",
       "                       [[ 6.3067e-02, -9.6567e-02, -8.9004e-02],\n",
       "                        [-5.3989e-02,  6.7611e-02,  7.0680e-02],\n",
       "                        [-7.1991e-02,  2.0100e-02, -5.5854e-02]]],\n",
       "              \n",
       "              \n",
       "                      [[[-4.8926e-02,  9.0907e-02,  5.0914e-02],\n",
       "                        [-2.8828e-02,  1.5516e-02,  2.0424e-02],\n",
       "                        [ 2.4691e-02, -3.6079e-02, -6.2074e-02]],\n",
       "              \n",
       "                       [[ 6.9788e-02,  1.4164e-02,  4.4119e-02],\n",
       "                        [-3.9922e-02,  5.1057e-02,  7.6713e-02],\n",
       "                        [ 6.4107e-02,  2.8660e-02,  1.0371e-01]],\n",
       "              \n",
       "                       [[-2.3053e-04,  2.2441e-02,  1.0015e-01],\n",
       "                        [ 1.0245e-01, -4.4506e-02,  9.4953e-02],\n",
       "                        [ 3.8902e-02, -1.1799e-02,  9.2038e-02]],\n",
       "              \n",
       "                       [[-5.4605e-02,  6.8490e-02,  1.0445e-01],\n",
       "                        [-7.2701e-02, -6.2201e-02, -1.0445e-01],\n",
       "                        [-1.8970e-02, -9.5733e-02, -3.5304e-02]],\n",
       "              \n",
       "                       [[ 3.2002e-02,  7.4511e-02,  5.8717e-02],\n",
       "                        [ 5.8511e-02,  4.3730e-02, -6.5378e-02],\n",
       "                        [-8.3694e-02,  4.3696e-03,  1.0009e-01]],\n",
       "              \n",
       "                       [[ 5.9351e-03, -9.0662e-03, -7.1545e-02],\n",
       "                        [-5.2266e-02, -8.1256e-02,  8.4398e-02],\n",
       "                        [-1.7174e-02, -9.3119e-02,  1.1308e-02]],\n",
       "              \n",
       "                       [[ 7.6494e-03, -1.3023e-02,  3.7733e-02],\n",
       "                        [ 5.6687e-02, -9.9128e-02, -8.0753e-02],\n",
       "                        [-5.0639e-03, -9.7729e-02, -9.5750e-02]],\n",
       "              \n",
       "                       [[ 9.3067e-02, -8.0174e-03, -5.2113e-02],\n",
       "                        [-3.6157e-02, -8.2295e-02,  8.2258e-02],\n",
       "                        [-2.2857e-02, -5.9265e-02, -7.9944e-02]],\n",
       "              \n",
       "                       [[ 6.1611e-02, -1.4571e-02, -1.1074e-02],\n",
       "                        [-2.7473e-02, -5.0883e-02,  1.8751e-02],\n",
       "                        [ 8.1099e-02, -6.1093e-02,  5.0504e-03]],\n",
       "              \n",
       "                       [[-8.0165e-02, -4.9426e-02,  9.2525e-02],\n",
       "                        [ 1.1052e-03,  1.0154e-01, -1.8468e-02],\n",
       "                        [-5.7453e-02, -6.2981e-02,  9.3426e-02]]],\n",
       "              \n",
       "              \n",
       "                      [[[-8.1058e-02,  5.5318e-02,  2.6203e-02],\n",
       "                        [ 3.1107e-02,  5.9476e-02, -2.7577e-02],\n",
       "                        [ 6.5223e-02, -8.3982e-02, -3.7087e-02]],\n",
       "              \n",
       "                       [[ 7.7164e-02,  3.1283e-02, -1.4038e-02],\n",
       "                        [-2.4616e-02, -6.4364e-02,  6.4098e-02],\n",
       "                        [-3.3520e-03, -3.5664e-03,  2.4929e-02]],\n",
       "              \n",
       "                       [[ 7.7787e-02, -5.3778e-02, -3.6303e-02],\n",
       "                        [ 7.1429e-02,  5.9532e-02, -5.1855e-02],\n",
       "                        [-1.0428e-01,  1.9555e-02,  5.5434e-02]],\n",
       "              \n",
       "                       [[ 2.5178e-02,  7.4768e-02, -8.3640e-02],\n",
       "                        [ 5.3156e-02, -6.5531e-02,  5.9325e-02],\n",
       "                        [ 7.8394e-02,  3.3385e-02,  8.5284e-02]],\n",
       "              \n",
       "                       [[-6.9481e-02, -9.4275e-02, -1.0135e-01],\n",
       "                        [ 6.6179e-02,  3.6926e-02, -7.7188e-02],\n",
       "                        [ 5.1048e-02,  9.6177e-02, -1.0394e-01]],\n",
       "              \n",
       "                       [[ 7.6466e-02,  1.6167e-02,  9.8053e-03],\n",
       "                        [ 9.4847e-02,  9.5458e-02,  4.4414e-02],\n",
       "                        [ 8.3288e-02,  4.3853e-02,  1.7176e-02]],\n",
       "              \n",
       "                       [[-9.2656e-02,  1.9689e-02, -7.4993e-02],\n",
       "                        [ 3.2452e-02,  1.8598e-02,  2.3681e-03],\n",
       "                        [-7.2071e-02, -6.3899e-02,  7.7912e-02]],\n",
       "              \n",
       "                       [[ 5.1336e-02,  5.5576e-02, -3.1410e-02],\n",
       "                        [-1.8151e-02, -2.7014e-02,  7.2489e-02],\n",
       "                        [-4.5504e-02,  6.6394e-02,  7.2679e-02]],\n",
       "              \n",
       "                       [[-9.6403e-02,  6.4369e-04, -2.0076e-02],\n",
       "                        [-5.8273e-02,  4.5507e-02, -1.2807e-02],\n",
       "                        [ 9.2287e-02, -6.5976e-02,  4.8976e-02]],\n",
       "              \n",
       "                       [[-8.9998e-02, -5.2833e-02,  7.1903e-03],\n",
       "                        [ 8.3283e-02,  5.5521e-02, -8.6550e-02],\n",
       "                        [ 1.1676e-02, -6.2138e-02,  4.5674e-03]]]], device='cuda:0')),\n",
       "             ('block_2.0.bias',\n",
       "              tensor([-0.0878, -0.0309,  0.0723, -0.0967, -0.1005,  0.0192,  0.0144, -0.0193,\n",
       "                       0.0920, -0.0635], device='cuda:0')),\n",
       "             ('block_2.2.weight',\n",
       "              tensor([[[[-6.3992e-02, -7.8791e-02, -1.9619e-02],\n",
       "                        [-2.6901e-02,  6.5222e-02, -5.9186e-03],\n",
       "                        [ 3.3663e-02, -4.3804e-02,  8.5507e-02]],\n",
       "              \n",
       "                       [[ 8.8862e-02, -9.4401e-02, -2.7090e-02],\n",
       "                        [-8.9439e-02,  4.4781e-02, -9.2094e-02],\n",
       "                        [-4.9839e-02,  1.0532e-01, -1.0066e-01]],\n",
       "              \n",
       "                       [[ 7.7771e-02,  8.9049e-03,  8.4289e-02],\n",
       "                        [-5.3494e-02,  6.9236e-02,  1.2718e-02],\n",
       "                        [ 8.1073e-03,  7.1945e-02, -1.0019e-01]],\n",
       "              \n",
       "                       [[-8.4902e-02,  1.0180e-01, -6.3298e-02],\n",
       "                        [-7.5980e-02, -5.1539e-03, -3.3742e-02],\n",
       "                        [-1.4421e-02, -7.0623e-02,  3.8034e-02]],\n",
       "              \n",
       "                       [[-9.0703e-02,  8.5374e-03,  6.1510e-02],\n",
       "                        [ 2.0253e-02,  1.4006e-02,  1.5418e-02],\n",
       "                        [-3.0880e-02, -2.0080e-02, -4.4450e-02]],\n",
       "              \n",
       "                       [[-7.1207e-02, -5.5810e-02,  1.0420e-01],\n",
       "                        [-1.7641e-02,  3.6924e-02,  7.2896e-02],\n",
       "                        [-8.2343e-03, -5.6707e-02, -7.1419e-02]],\n",
       "              \n",
       "                       [[-3.8833e-02,  3.7624e-02, -8.8771e-02],\n",
       "                        [-1.2870e-02,  4.0096e-02,  8.5999e-02],\n",
       "                        [ 3.1721e-02,  2.0846e-02,  7.2162e-02]],\n",
       "              \n",
       "                       [[ 4.8708e-02,  3.5661e-02, -3.2682e-02],\n",
       "                        [-8.4528e-02, -2.2769e-02, -1.9117e-02],\n",
       "                        [ 7.7410e-03, -1.1593e-02,  4.2616e-02]],\n",
       "              \n",
       "                       [[ 7.0050e-02, -4.2735e-02, -1.0002e-01],\n",
       "                        [-5.4081e-02, -5.0436e-02,  5.9750e-02],\n",
       "                        [-6.7994e-02, -9.9145e-03, -2.2340e-02]],\n",
       "              \n",
       "                       [[-6.3976e-02,  4.7780e-02, -4.3909e-02],\n",
       "                        [-5.4531e-03, -7.4112e-02, -1.0632e-02],\n",
       "                        [ 1.4977e-02, -4.2894e-03, -3.9386e-02]]],\n",
       "              \n",
       "              \n",
       "                      [[[ 3.1315e-02, -2.7311e-02, -5.8439e-02],\n",
       "                        [-7.7732e-02, -2.2329e-02, -9.9578e-02],\n",
       "                        [ 8.7492e-02, -5.0357e-02, -4.3684e-02]],\n",
       "              \n",
       "                       [[ 9.7439e-03,  2.7326e-02, -9.9393e-03],\n",
       "                        [ 7.2313e-02, -6.1448e-02,  3.7777e-02],\n",
       "                        [-2.3773e-04, -8.5747e-02, -4.0824e-02]],\n",
       "              \n",
       "                       [[ 2.6825e-02,  2.0138e-02,  7.6647e-02],\n",
       "                        [ 7.0518e-02, -5.7493e-02, -4.5013e-02],\n",
       "                        [-2.2351e-02, -7.5517e-02, -2.8459e-02]],\n",
       "              \n",
       "                       [[-8.6258e-02,  4.0092e-02,  7.4583e-02],\n",
       "                        [ 8.3459e-03, -7.5460e-02, -7.9827e-02],\n",
       "                        [-4.1036e-02,  3.0659e-02,  2.5711e-03]],\n",
       "              \n",
       "                       [[ 1.9166e-02,  9.9346e-02,  4.8956e-02],\n",
       "                        [ 2.2665e-02, -2.1327e-02,  4.9864e-02],\n",
       "                        [ 3.8563e-02, -9.4879e-02, -6.2266e-02]],\n",
       "              \n",
       "                       [[ 3.5381e-03,  3.9997e-02,  5.1282e-02],\n",
       "                        [-6.2748e-02, -1.0458e-01, -5.4909e-03],\n",
       "                        [-1.2050e-02,  3.0588e-02, -2.8988e-02]],\n",
       "              \n",
       "                       [[ 8.0588e-02,  7.0333e-03,  7.6975e-02],\n",
       "                        [-7.3398e-02,  4.2167e-02,  1.2560e-02],\n",
       "                        [-5.2720e-02,  5.2256e-02, -1.0372e-01]],\n",
       "              \n",
       "                       [[ 8.5220e-02,  8.4947e-03,  1.0178e-02],\n",
       "                        [ 4.8746e-02,  8.7503e-03,  4.5184e-02],\n",
       "                        [ 6.7063e-02, -8.2268e-02,  6.9735e-02]],\n",
       "              \n",
       "                       [[-1.5784e-02, -2.4513e-02,  2.1217e-02],\n",
       "                        [ 8.2446e-02, -5.7302e-02, -7.1039e-02],\n",
       "                        [ 6.5418e-02, -4.9507e-02,  3.3937e-02]],\n",
       "              \n",
       "                       [[-1.5530e-02,  2.9014e-02,  8.0439e-02],\n",
       "                        [-5.3421e-02, -5.1151e-02,  5.1716e-02],\n",
       "                        [ 5.7714e-03, -1.1601e-02, -9.2590e-02]]],\n",
       "              \n",
       "              \n",
       "                      [[[ 8.9309e-02, -3.9919e-03, -1.9415e-02],\n",
       "                        [-4.3269e-02, -2.0801e-02,  5.1233e-02],\n",
       "                        [-2.4227e-03,  9.0147e-02, -6.0858e-03]],\n",
       "              \n",
       "                       [[-1.5122e-02,  5.9498e-02, -2.7275e-03],\n",
       "                        [-2.1039e-02,  3.5231e-02,  8.3129e-02],\n",
       "                        [ 2.6305e-02,  7.3398e-02,  6.8309e-02]],\n",
       "              \n",
       "                       [[ 2.9810e-02,  3.6650e-02,  3.4014e-02],\n",
       "                        [ 1.0934e-02,  8.9675e-02,  9.7308e-02],\n",
       "                        [ 3.7524e-02, -5.2640e-03,  9.4509e-02]],\n",
       "              \n",
       "                       [[-8.2042e-02,  7.7453e-02,  5.5849e-02],\n",
       "                        [ 6.7687e-02, -8.0992e-03, -7.8646e-02],\n",
       "                        [ 7.5193e-02, -4.6091e-02,  2.7734e-02]],\n",
       "              \n",
       "                       [[ 5.9719e-02, -9.8508e-02,  6.9954e-03],\n",
       "                        [-3.7444e-02,  7.4815e-02, -6.7114e-02],\n",
       "                        [ 6.4001e-02,  6.5730e-02,  5.8156e-02]],\n",
       "              \n",
       "                       [[ 1.0119e-01,  1.5964e-02, -9.5541e-02],\n",
       "                        [ 7.5248e-02,  9.6499e-03,  2.0918e-03],\n",
       "                        [-1.0041e-01, -2.3691e-02, -5.1162e-02]],\n",
       "              \n",
       "                       [[ 1.0324e-01,  7.5054e-02,  7.8634e-02],\n",
       "                        [ 7.2188e-02, -6.5340e-02, -4.5270e-02],\n",
       "                        [-4.1252e-02, -4.2257e-02,  8.2054e-02]],\n",
       "              \n",
       "                       [[ 3.5815e-02,  8.4470e-02, -4.9309e-03],\n",
       "                        [-9.3965e-02, -3.0582e-02,  7.4081e-02],\n",
       "                        [ 6.4174e-02,  3.2632e-02, -3.0919e-02]],\n",
       "              \n",
       "                       [[-9.8386e-02, -5.6639e-02,  5.4958e-02],\n",
       "                        [-4.2518e-02,  5.0421e-02,  2.8781e-02],\n",
       "                        [-4.0486e-02,  6.4202e-02, -3.3871e-02]],\n",
       "              \n",
       "                       [[-3.5020e-03, -4.0152e-02, -9.9988e-02],\n",
       "                        [ 1.6996e-02,  3.0460e-02, -5.3072e-02],\n",
       "                        [ 6.4663e-02, -9.4558e-02, -1.0161e-01]]],\n",
       "              \n",
       "              \n",
       "                      [[[-6.5106e-02, -3.6430e-02, -1.1707e-02],\n",
       "                        [-2.0370e-02,  4.8108e-02, -9.2510e-02],\n",
       "                        [ 1.5521e-02,  1.8254e-03,  2.7842e-02]],\n",
       "              \n",
       "                       [[ 1.0479e-01,  6.4874e-02, -5.8366e-02],\n",
       "                        [-8.6378e-02, -2.5520e-02, -5.2876e-02],\n",
       "                        [ 3.6820e-02,  9.6628e-04,  8.4783e-02]],\n",
       "              \n",
       "                       [[ 4.1405e-02, -1.9382e-02,  3.6229e-03],\n",
       "                        [ 2.5244e-02, -1.3080e-02,  8.5058e-02],\n",
       "                        [-8.2420e-02,  5.1377e-02, -6.7192e-02]],\n",
       "              \n",
       "                       [[-9.2347e-02, -2.1640e-02,  5.1366e-02],\n",
       "                        [ 7.4478e-02,  2.6452e-02, -9.1104e-03],\n",
       "                        [-5.9092e-03, -4.2731e-02, -9.4592e-03]],\n",
       "              \n",
       "                       [[-7.2831e-03,  8.9699e-02,  6.1690e-02],\n",
       "                        [-8.4351e-02,  4.3605e-04, -6.4834e-02],\n",
       "                        [-1.6733e-02, -8.3776e-02,  2.7402e-02]],\n",
       "              \n",
       "                       [[-7.6008e-02,  1.0406e-01,  7.9605e-02],\n",
       "                        [-7.2559e-02, -9.9239e-02,  4.1128e-03],\n",
       "                        [-2.9425e-02,  3.0945e-02, -7.1353e-02]],\n",
       "              \n",
       "                       [[ 4.3148e-02, -9.1047e-02, -5.5632e-02],\n",
       "                        [-5.5414e-02,  5.1007e-02, -2.7597e-03],\n",
       "                        [-1.0130e-01, -6.0201e-02, -4.8781e-02]],\n",
       "              \n",
       "                       [[-9.7802e-02,  1.3497e-02,  3.7561e-02],\n",
       "                        [-1.9340e-02, -4.1947e-02, -6.3926e-04],\n",
       "                        [-8.3725e-02, -6.4184e-02, -2.4040e-03]],\n",
       "              \n",
       "                       [[ 9.3643e-02, -3.2414e-02,  5.2247e-02],\n",
       "                        [-4.1484e-02, -2.8060e-02, -1.0034e-01],\n",
       "                        [ 8.7330e-02,  1.0264e-01, -2.2139e-03]],\n",
       "              \n",
       "                       [[ 6.6974e-02,  8.6219e-02,  5.2359e-02],\n",
       "                        [ 5.4288e-02, -1.0035e-01, -9.9050e-02],\n",
       "                        [-8.0906e-02,  3.2970e-02, -9.1177e-02]]],\n",
       "              \n",
       "              \n",
       "                      [[[-8.0464e-02, -5.1092e-02, -9.7154e-02],\n",
       "                        [ 1.4203e-04,  1.5207e-02, -6.1686e-02],\n",
       "                        [ 6.9018e-02, -4.0018e-02, -2.9676e-02]],\n",
       "              \n",
       "                       [[ 8.0309e-02,  9.0499e-02, -1.2093e-02],\n",
       "                        [-7.5671e-02, -5.2881e-02,  1.3423e-02],\n",
       "                        [ 6.1790e-02,  5.2477e-02, -4.6547e-02]],\n",
       "              \n",
       "                       [[-9.9650e-02, -9.2249e-02, -3.3537e-02],\n",
       "                        [ 1.3223e-03, -4.7347e-02, -8.3348e-02],\n",
       "                        [ 1.1109e-02, -8.3668e-02, -8.0946e-02]],\n",
       "              \n",
       "                       [[-8.5692e-02, -2.8563e-02,  9.3104e-02],\n",
       "                        [ 4.1207e-02, -1.2498e-02,  2.1694e-02],\n",
       "                        [ 4.1975e-02,  6.1414e-04, -8.5020e-02]],\n",
       "              \n",
       "                       [[-6.4944e-02, -7.1610e-02, -2.6766e-03],\n",
       "                        [-9.6492e-02, -1.9166e-02, -3.8545e-02],\n",
       "                        [ 1.0345e-01,  8.5679e-02,  6.1227e-02]],\n",
       "              \n",
       "                       [[ 5.9116e-03, -3.4129e-02,  2.6887e-02],\n",
       "                        [-7.2830e-02, -4.4957e-02, -2.1175e-02],\n",
       "                        [-2.4766e-02, -9.9854e-02,  4.1903e-02]],\n",
       "              \n",
       "                       [[ 8.6803e-02, -5.8141e-02,  2.8415e-02],\n",
       "                        [-1.2225e-02, -3.8445e-03,  6.1443e-03],\n",
       "                        [ 9.1346e-02,  1.4124e-02, -6.6690e-02]],\n",
       "              \n",
       "                       [[-3.7917e-02,  5.1495e-02,  3.2893e-02],\n",
       "                        [ 2.0487e-03, -1.3912e-02, -4.1012e-02],\n",
       "                        [-3.7413e-02, -5.5602e-02,  1.7273e-02]],\n",
       "              \n",
       "                       [[ 2.9603e-02,  8.0717e-02, -2.3813e-02],\n",
       "                        [ 7.5461e-03,  6.8125e-02,  4.5852e-02],\n",
       "                        [ 1.3544e-02,  3.2390e-02,  5.4714e-03]],\n",
       "              \n",
       "                       [[-9.0419e-02,  4.0636e-03, -2.3040e-02],\n",
       "                        [ 9.5123e-02,  9.5145e-02,  2.0912e-02],\n",
       "                        [ 9.4215e-02, -5.4288e-02,  9.1619e-02]]],\n",
       "              \n",
       "              \n",
       "                      [[[ 9.0756e-02, -4.0288e-03, -8.4592e-02],\n",
       "                        [-3.4015e-02, -2.8189e-02,  1.7411e-03],\n",
       "                        [-9.5569e-02,  1.9535e-02, -4.3839e-02]],\n",
       "              \n",
       "                       [[-2.6989e-02, -5.4443e-02, -2.2255e-02],\n",
       "                        [-9.7896e-02, -5.5885e-02,  9.7108e-03],\n",
       "                        [ 6.9072e-02,  9.5790e-02, -7.9737e-02]],\n",
       "              \n",
       "                       [[ 4.4264e-02, -5.9419e-02, -8.1498e-02],\n",
       "                        [-4.6417e-03, -6.0468e-02, -9.0783e-02],\n",
       "                        [-9.8509e-02, -7.0556e-02,  8.6619e-02]],\n",
       "              \n",
       "                       [[ 5.8788e-02, -4.1726e-02, -7.0553e-02],\n",
       "                        [-8.1085e-02, -6.2246e-02, -4.3376e-02],\n",
       "                        [ 6.3308e-02,  3.4496e-02, -4.0622e-02]],\n",
       "              \n",
       "                       [[ 7.2567e-02, -6.5484e-02, -8.5876e-02],\n",
       "                        [ 2.3006e-02, -5.8123e-02,  2.9987e-02],\n",
       "                        [ 8.9306e-02, -4.9849e-02, -7.3556e-02]],\n",
       "              \n",
       "                       [[ 3.9676e-02, -9.5200e-02,  9.4044e-02],\n",
       "                        [-4.9780e-02,  5.0961e-02, -8.3818e-02],\n",
       "                        [-7.1348e-02,  1.1611e-02,  3.7463e-02]],\n",
       "              \n",
       "                       [[ 8.1734e-02,  8.8158e-02, -6.0623e-03],\n",
       "                        [-1.3552e-02,  1.7424e-02, -2.4486e-02],\n",
       "                        [ 3.5882e-03, -9.9828e-02, -8.6531e-02]],\n",
       "              \n",
       "                       [[ 7.2233e-02, -6.1597e-02,  8.3008e-02],\n",
       "                        [ 1.1568e-02,  2.5676e-02,  9.5804e-02],\n",
       "                        [-5.8628e-02, -1.6640e-02,  1.8675e-02]],\n",
       "              \n",
       "                       [[ 3.6012e-02, -1.0259e-01,  3.7464e-02],\n",
       "                        [-6.2163e-02,  1.3846e-02,  7.1315e-02],\n",
       "                        [-1.0500e-02, -3.3346e-03, -7.8757e-03]],\n",
       "              \n",
       "                       [[ 8.7962e-02,  5.9907e-02,  1.7727e-02],\n",
       "                        [-6.3437e-02, -5.7241e-02,  8.3964e-02],\n",
       "                        [ 7.5834e-02,  6.1033e-02, -8.2189e-02]]],\n",
       "              \n",
       "              \n",
       "                      [[[ 8.2092e-02, -1.0076e-02,  7.7661e-02],\n",
       "                        [ 9.1553e-02,  1.1554e-02, -4.3863e-02],\n",
       "                        [ 9.9153e-02, -5.4931e-02,  6.8876e-02]],\n",
       "              \n",
       "                       [[-1.0108e-01, -3.3153e-02, -9.1902e-02],\n",
       "                        [-4.7284e-02,  4.4759e-02, -7.5529e-02],\n",
       "                        [-9.1158e-02,  7.5371e-02,  5.6270e-02]],\n",
       "              \n",
       "                       [[-1.1527e-03, -7.4309e-02, -2.7927e-02],\n",
       "                        [-3.4129e-02,  6.5100e-02, -3.4478e-02],\n",
       "                        [-3.0360e-02, -7.4720e-02, -4.9646e-02]],\n",
       "              \n",
       "                       [[ 5.7074e-02,  6.7914e-02,  1.5315e-02],\n",
       "                        [-3.9549e-02,  1.0124e-01,  2.0806e-02],\n",
       "                        [-4.0688e-02, -3.6535e-02, -1.4752e-02]],\n",
       "              \n",
       "                       [[ 4.9974e-02,  3.8555e-02,  7.6418e-02],\n",
       "                        [-4.7494e-03,  8.7183e-02, -4.2816e-02],\n",
       "                        [-4.8547e-02, -3.8927e-02, -9.8896e-02]],\n",
       "              \n",
       "                       [[-6.9195e-02, -9.5382e-02, -6.2294e-03],\n",
       "                        [ 9.9374e-04, -2.7358e-02, -7.2035e-02],\n",
       "                        [ 9.5637e-02, -3.4926e-02,  5.0233e-02]],\n",
       "              \n",
       "                       [[ 7.3408e-02, -6.9292e-02, -1.3179e-02],\n",
       "                        [ 6.0923e-02,  1.0218e-01, -1.3299e-02],\n",
       "                        [ 7.6382e-02, -8.2732e-02, -6.8489e-02]],\n",
       "              \n",
       "                       [[ 8.6682e-02, -9.9801e-03,  1.0414e-01],\n",
       "                        [ 7.6651e-03, -4.3714e-02,  1.0011e-01],\n",
       "                        [ 9.2179e-02,  9.7826e-03, -6.3900e-02]],\n",
       "              \n",
       "                       [[-4.5639e-03, -5.0693e-02,  7.6810e-02],\n",
       "                        [ 4.8829e-03,  2.2191e-02,  6.3927e-02],\n",
       "                        [ 3.4916e-02, -6.5803e-02,  8.7566e-02]],\n",
       "              \n",
       "                       [[ 6.4758e-02, -6.5073e-02,  7.9700e-02],\n",
       "                        [ 2.9905e-02, -2.0750e-02, -7.5385e-02],\n",
       "                        [-1.7490e-02, -1.0335e-01,  6.0163e-02]]],\n",
       "              \n",
       "              \n",
       "                      [[[ 7.6343e-02, -3.0347e-02,  9.7720e-02],\n",
       "                        [-3.9032e-02,  1.8051e-02, -7.3459e-02],\n",
       "                        [-4.4565e-03,  4.2610e-02,  4.5403e-02]],\n",
       "              \n",
       "                       [[-3.5346e-03, -5.3154e-02,  7.3680e-02],\n",
       "                        [ 6.9788e-02,  1.6916e-02, -4.8475e-02],\n",
       "                        [ 2.2349e-02,  2.8186e-04,  9.6302e-02]],\n",
       "              \n",
       "                       [[ 1.5621e-02,  8.1301e-03,  7.2057e-03],\n",
       "                        [ 5.6079e-02, -1.3024e-03,  9.0351e-02],\n",
       "                        [ 5.4917e-02, -7.9650e-02, -1.2070e-06]],\n",
       "              \n",
       "                       [[-8.9472e-02, -8.0934e-02,  2.0480e-02],\n",
       "                        [ 2.3687e-02, -9.2246e-03,  1.0019e-01],\n",
       "                        [-5.6627e-02, -4.4176e-02, -1.6881e-02]],\n",
       "              \n",
       "                       [[ 6.3911e-04, -8.9284e-03,  9.4909e-02],\n",
       "                        [-4.4519e-02, -5.5137e-02,  9.0599e-03],\n",
       "                        [ 7.9171e-02,  2.5019e-02,  5.6787e-02]],\n",
       "              \n",
       "                       [[ 2.0406e-02,  8.9839e-02,  6.3311e-02],\n",
       "                        [ 7.5428e-02, -1.4198e-02, -8.7268e-02],\n",
       "                        [-5.0002e-02,  3.5910e-02,  7.3950e-02]],\n",
       "              \n",
       "                       [[-4.1184e-02,  8.7218e-02,  1.5150e-02],\n",
       "                        [ 4.1869e-04,  4.1093e-03, -1.8623e-02],\n",
       "                        [ 9.8683e-02,  4.5784e-03,  6.4564e-02]],\n",
       "              \n",
       "                       [[-8.8967e-02, -5.4309e-02,  1.1852e-02],\n",
       "                        [ 8.4169e-02,  5.0184e-02,  2.0076e-02],\n",
       "                        [-1.0414e-01,  1.9816e-03, -6.9581e-02]],\n",
       "              \n",
       "                       [[-9.0006e-02,  1.4414e-02, -6.6693e-02],\n",
       "                        [ 9.5674e-02, -5.7294e-02,  3.3970e-02],\n",
       "                        [ 6.1871e-02, -8.1928e-02,  5.3946e-02]],\n",
       "              \n",
       "                       [[-1.4114e-02,  5.4619e-02,  1.0201e-01],\n",
       "                        [-4.4922e-02, -4.5653e-02,  8.3753e-02],\n",
       "                        [ 1.1722e-02, -1.0513e-02,  7.9971e-02]]],\n",
       "              \n",
       "              \n",
       "                      [[[-5.0928e-02, -5.2047e-03,  7.2403e-02],\n",
       "                        [ 4.1195e-02, -6.8180e-02,  2.7398e-02],\n",
       "                        [-8.0368e-02, -5.7245e-02,  6.7779e-02]],\n",
       "              \n",
       "                       [[-2.8093e-02, -5.3691e-02,  7.4717e-03],\n",
       "                        [ 2.5759e-02, -6.5524e-02, -7.1084e-02],\n",
       "                        [-1.0209e-01,  2.7236e-02, -6.8013e-02]],\n",
       "              \n",
       "                       [[ 8.0331e-03, -2.3576e-02, -6.8923e-02],\n",
       "                        [-3.3636e-02, -8.1027e-02, -5.5797e-02],\n",
       "                        [-3.2857e-03, -9.0116e-02, -9.2447e-02]],\n",
       "              \n",
       "                       [[ 7.8958e-02,  9.9188e-03, -4.6618e-02],\n",
       "                        [-3.5047e-03,  7.8168e-02, -8.7939e-02],\n",
       "                        [-5.5886e-02, -7.6226e-02, -7.6634e-03]],\n",
       "              \n",
       "                       [[-3.6274e-03, -8.2146e-02,  7.3163e-02],\n",
       "                        [-8.0946e-02,  9.8414e-02, -7.2560e-02],\n",
       "                        [-1.4446e-02,  1.9710e-02, -4.6852e-02]],\n",
       "              \n",
       "                       [[ 9.6939e-02, -7.2673e-02, -5.8427e-03],\n",
       "                        [-7.7398e-02,  2.9261e-02,  8.9871e-02],\n",
       "                        [ 9.7776e-02,  1.2514e-02, -5.2773e-02]],\n",
       "              \n",
       "                       [[ 1.0244e-01,  7.8667e-03,  7.1317e-02],\n",
       "                        [-5.4751e-02, -4.8920e-02, -8.7504e-02],\n",
       "                        [ 9.6990e-02,  1.7486e-02, -7.5704e-02]],\n",
       "              \n",
       "                       [[ 9.0535e-03, -4.5211e-02,  5.2659e-03],\n",
       "                        [ 3.4988e-02, -5.2308e-02,  1.8394e-02],\n",
       "                        [-6.6553e-02,  2.0312e-02, -1.0178e-01]],\n",
       "              \n",
       "                       [[ 1.6797e-02,  1.0473e-01,  9.7094e-02],\n",
       "                        [ 3.8451e-02,  7.7563e-02,  1.0248e-01],\n",
       "                        [ 2.9870e-02,  3.5156e-02,  1.3707e-02]],\n",
       "              \n",
       "                       [[ 9.3322e-02,  9.0551e-02, -4.9570e-02],\n",
       "                        [-4.3333e-03, -5.3110e-02,  3.7824e-02],\n",
       "                        [-1.0214e-01,  3.7301e-02, -2.8929e-02]]],\n",
       "              \n",
       "              \n",
       "                      [[[ 3.8227e-02,  3.2899e-02, -5.2454e-02],\n",
       "                        [ 5.4687e-02,  4.4762e-02, -8.9602e-02],\n",
       "                        [ 1.0517e-01,  9.0731e-02,  6.5584e-02]],\n",
       "              \n",
       "                       [[-1.0699e-02,  3.7345e-02, -5.7028e-02],\n",
       "                        [-3.5818e-02,  4.9749e-02,  4.6925e-02],\n",
       "                        [ 4.1741e-02, -1.0053e-01,  8.7350e-02]],\n",
       "              \n",
       "                       [[-4.4028e-02,  9.1223e-02,  8.6852e-02],\n",
       "                        [ 3.9070e-02,  1.0502e-01,  6.0528e-02],\n",
       "                        [ 6.1821e-02, -3.5794e-02,  9.7766e-02]],\n",
       "              \n",
       "                       [[ 2.7627e-02,  6.2280e-02, -2.3834e-02],\n",
       "                        [ 7.6340e-02,  9.3509e-02, -8.0770e-02],\n",
       "                        [ 8.6415e-02, -6.9664e-02, -7.2571e-02]],\n",
       "              \n",
       "                       [[-8.8089e-02,  3.0459e-02, -7.9144e-02],\n",
       "                        [-3.9680e-02, -5.2988e-02,  2.8172e-02],\n",
       "                        [-1.0349e-01, -4.8324e-02,  7.7112e-04]],\n",
       "              \n",
       "                       [[ 9.4660e-03, -4.7605e-02,  3.7764e-02],\n",
       "                        [-6.9544e-02, -8.9270e-02, -1.4986e-02],\n",
       "                        [-5.6989e-02,  6.6443e-02, -7.2049e-02]],\n",
       "              \n",
       "                       [[-8.8494e-03,  4.3782e-02, -9.2311e-02],\n",
       "                        [ 8.1599e-02, -4.7895e-02, -2.8684e-02],\n",
       "                        [-6.4480e-02, -3.9279e-02, -4.0645e-02]],\n",
       "              \n",
       "                       [[-9.3801e-02,  3.6019e-02, -3.3768e-04],\n",
       "                        [ 1.0311e-01,  7.1117e-02,  9.1699e-02],\n",
       "                        [ 3.1014e-02,  5.5388e-02,  9.8704e-02]],\n",
       "              \n",
       "                       [[ 8.6545e-02, -8.0996e-02, -2.3636e-02],\n",
       "                        [-1.0166e-01,  3.9877e-03, -3.7229e-02],\n",
       "                        [ 9.1486e-02,  1.6666e-02,  1.1601e-03]],\n",
       "              \n",
       "                       [[-7.6248e-02, -8.2718e-02,  1.6594e-02],\n",
       "                        [-5.2376e-02, -4.8409e-02,  7.3938e-02],\n",
       "                        [-5.4952e-02, -4.6918e-02,  8.0934e-02]]]], device='cuda:0')),\n",
       "             ('block_2.2.bias',\n",
       "              tensor([ 0.0412, -0.0599,  0.0319,  0.0531, -0.0936,  0.0197,  0.0241, -0.0041,\n",
       "                       0.1011, -0.0697], device='cuda:0')),\n",
       "             ('classifier.1.weight',\n",
       "              tensor([[ 0.0245, -0.0240, -0.0387,  ...,  0.0094, -0.0015, -0.0225],\n",
       "                      [ 0.0228,  0.0067, -0.0439,  ..., -0.0302,  0.0368,  0.0293],\n",
       "                      [ 0.0303,  0.0347, -0.0211,  ...,  0.0207, -0.0423, -0.0240],\n",
       "                      ...,\n",
       "                      [-0.0359, -0.0343,  0.0166,  ...,  0.0324,  0.0113, -0.0143],\n",
       "                      [-0.0294, -0.0316,  0.0251,  ..., -0.0056,  0.0300, -0.0396],\n",
       "                      [-0.0246, -0.0035, -0.0046,  ..., -0.0146, -0.0358,  0.0175]],\n",
       "                     device='cuda:0')),\n",
       "             ('classifier.1.bias',\n",
       "              tensor([ 0.0320, -0.0445,  0.0246, -0.0357, -0.0442,  0.0156, -0.0010, -0.0277,\n",
       "                       0.0404,  0.0037], device='cuda:0'))])"
      ]
     },
     "execution_count": 77,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model_2.state_dict()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6478cc5a-7b33-425d-9ab3-6d40168a1aee",
   "metadata": {
    "id": "6478cc5a-7b33-425d-9ab3-6d40168a1aee"
   },
   "source": [
    "### 7.1 Stepping through `nn.Conv2d()`\n",
    "\n",
    "We could start using our model above and see what happens but let's first step through the two new layers we've added:\n",
    "* [`nn.Conv2d()`](https://pytorch.org/docs/stable/generated/torch.nn.Conv2d.html), also known as a convolutional layer.\n",
    "* [`nn.MaxPool2d()`](https://pytorch.org/docs/stable/generated/torch.nn.MaxPool2d.html), also known as a max pooling layer.\n",
    "\n",
    "> **Question:** What does the \"2d\" in `nn.Conv2d()` stand for?\n",
    ">\n",
    "> The 2d is for 2-dimensional data. As in, our images have two dimensions: height and width. Yes, there's color channel dimension but each of the color channel dimensions have two dimensions too: height and width.\n",
    ">\n",
    "> For other dimensional data (such as 1D for text or 3D for 3D objects) there's also `nn.Conv1d()` and `nn.Conv3d()`. \n",
    "\n",
    "To test the layers out, let's create some toy data just like the data used on CNN Explainer."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "id": "058b01ac-3f6a-4472-bcbf-3377974e3254",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "058b01ac-3f6a-4472-bcbf-3377974e3254",
    "outputId": "c404a8dd-d804-4993-bc2b-e4fdbf02b62d"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Image batch shape: torch.Size([32, 3, 64, 64]) -> [batch_size, color_channels, height, width]\n",
      "Single image shape: torch.Size([3, 64, 64]) -> [color_channels, height, width]\n",
      "Single image pixel values:\n",
      "tensor([[[ 1.9269,  1.4873,  0.9007,  ...,  1.8446, -1.1845,  1.3835],\n",
      "         [ 1.4451,  0.8564,  2.2181,  ...,  0.3399,  0.7200,  0.4114],\n",
      "         [ 1.9312,  1.0119, -1.4364,  ..., -0.5558,  0.7043,  0.7099],\n",
      "         ...,\n",
      "         [-0.5610, -0.4830,  0.4770,  ..., -0.2713, -0.9537, -0.6737],\n",
      "         [ 0.3076, -0.1277,  0.0366,  ..., -2.0060,  0.2824, -0.8111],\n",
      "         [-1.5486,  0.0485, -0.7712,  ..., -0.1403,  0.9416, -0.0118]],\n",
      "\n",
      "        [[-0.5197,  1.8524,  1.8365,  ...,  0.8935, -1.5114, -0.8515],\n",
      "         [ 2.0818,  1.0677, -1.4277,  ...,  1.6612, -2.6223, -0.4319],\n",
      "         [-0.1010, -0.4388, -1.9775,  ...,  0.2106,  0.2536, -0.7318],\n",
      "         ...,\n",
      "         [ 0.2779,  0.7342, -0.3736,  ..., -0.4601,  0.1815,  0.1850],\n",
      "         [ 0.7205, -0.2833,  0.0937,  ..., -0.1002, -2.3609,  2.2465],\n",
      "         [-1.3242, -0.1973,  0.2920,  ...,  0.5409,  0.6940,  1.8563]],\n",
      "\n",
      "        [[-0.7978,  1.0261,  1.1465,  ...,  1.2134,  0.9354, -0.0780],\n",
      "         [-1.4647, -1.9571,  0.1017,  ..., -1.9986, -0.7409,  0.7011],\n",
      "         [-1.3938,  0.8466, -1.7191,  ..., -1.1867,  0.1320,  0.3407],\n",
      "         ...,\n",
      "         [ 0.8206, -0.3745,  1.2499,  ..., -0.0676,  0.0385,  0.6335],\n",
      "         [-0.5589, -0.3393,  0.2347,  ...,  2.1181,  2.4569,  1.3083],\n",
      "         [-0.4092,  1.5199,  0.2401,  ..., -0.2558,  0.7870,  0.9924]]])\n"
     ]
    }
   ],
   "source": [
    "torch.manual_seed(42)\n",
    "\n",
    "# Create sample batch of random numbers with same size as image batch\n",
    "images = torch.randn(size=(32, 3, 64, 64)) # [batch_size, color_channels, height, width]\n",
    "test_image = images[0] # get a single image for testing\n",
    "print(f\"Image batch shape: {images.shape} -> [batch_size, color_channels, height, width]\")\n",
    "print(f\"Single image shape: {test_image.shape} -> [color_channels, height, width]\") \n",
    "print(f\"Single image pixel values:\\n{test_image}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "bd3291c2-854e-4d0c-97b9-8bf46085fc43",
   "metadata": {
    "id": "bd3291c2-854e-4d0c-97b9-8bf46085fc43"
   },
   "source": [
    "Let's create an example `nn.Conv2d()` with various parameters:\n",
    "* `in_channels` (int) - Number of channels in the input image.\n",
    "* `out_channels` (int) - Number of channels produced by the convolution.\n",
    "* `kernel_size` (int or tuple) - Size of the convolving kernel/filter.\n",
    "* `stride` (int or tuple, optional) - How big of a step the convolving kernel takes at a time. Default: 1.\n",
    "* `padding` (int, tuple, str) - Padding added to all four sides of input. Default: 0.\n",
    "\n",
    "![example of going through the different parameters of a Conv2d layer](https://raw.githubusercontent.com/mrdbourke/pytorch-deep-learning/main/images/03-conv2d-layer.gif)\n",
    "\n",
    "*Example of what happens when you change the hyperparameters of a `nn.Conv2d()` layer.*"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "id": "ebd39562-1dad-40e3-90f5-750a5dac24e2",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "ebd39562-1dad-40e3-90f5-750a5dac24e2",
    "outputId": "b61154fb-c5f7-4c3f-c619-4bde6edb4d16"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tensor([[[ 1.5396,  0.0516,  0.6454,  ..., -0.3673,  0.8711,  0.4256],\n",
       "         [ 0.3662,  1.0114, -0.5997,  ...,  0.8983,  0.2809, -0.2741],\n",
       "         [ 1.2664, -1.4054,  0.3727,  ..., -0.3409,  1.2191, -0.0463],\n",
       "         ...,\n",
       "         [-0.1541,  0.5132, -0.3624,  ..., -0.2360, -0.4609, -0.0035],\n",
       "         [ 0.2981, -0.2432,  1.5012,  ..., -0.6289, -0.7283, -0.5767],\n",
       "         [-0.0386, -0.0781, -0.0388,  ...,  0.2842,  0.4228, -0.1802]],\n",
       "\n",
       "        [[-0.2840, -0.0319, -0.4455,  ..., -0.7956,  1.5599, -1.2449],\n",
       "         [ 0.2753, -0.1262, -0.6541,  ..., -0.2211,  0.1999, -0.8856],\n",
       "         [-0.5404, -1.5489,  0.0249,  ..., -0.5932, -1.0913, -0.3849],\n",
       "         ...,\n",
       "         [ 0.3870, -0.4064, -0.8236,  ...,  0.1734, -0.4330, -0.4951],\n",
       "         [-0.1984, -0.6386,  1.0263,  ..., -0.9401, -0.0585, -0.7833],\n",
       "         [-0.6306, -0.2052, -0.3694,  ..., -1.3248,  0.2456, -0.7134]],\n",
       "\n",
       "        [[ 0.4414,  0.5100,  0.4846,  ..., -0.8484,  0.2638,  1.1258],\n",
       "         [ 0.8117,  0.3191, -0.0157,  ...,  1.2686,  0.2319,  0.5003],\n",
       "         [ 0.3212,  0.0485, -0.2581,  ...,  0.2258,  0.2587, -0.8804],\n",
       "         ...,\n",
       "         [-0.1144, -0.1869,  0.0160,  ..., -0.8346,  0.0974,  0.8421],\n",
       "         [ 0.2941,  0.4417,  0.5866,  ..., -0.1224,  0.4814, -0.4799],\n",
       "         [ 0.6059, -0.0415, -0.2028,  ...,  0.1170,  0.2521, -0.4372]],\n",
       "\n",
       "        ...,\n",
       "\n",
       "        [[-0.2560, -0.0477,  0.6380,  ...,  0.6436,  0.7553, -0.7055],\n",
       "         [ 1.5595, -0.2209, -0.9486,  ..., -0.4876,  0.7754,  0.0750],\n",
       "         [-0.0797,  0.2471,  1.1300,  ...,  0.1505,  0.2354,  0.9576],\n",
       "         ...,\n",
       "         [ 1.1065,  0.6839,  1.2183,  ...,  0.3015, -0.1910, -0.1902],\n",
       "         [-0.3486, -0.7173, -0.3582,  ...,  0.4917,  0.7219,  0.1513],\n",
       "         [ 0.0119,  0.1017,  0.7839,  ..., -0.3752, -0.8127, -0.1257]],\n",
       "\n",
       "        [[ 0.3841,  1.1322,  0.1620,  ...,  0.7010,  0.0109,  0.6058],\n",
       "         [ 0.1664,  0.1873,  1.5924,  ...,  0.3733,  0.9096, -0.5399],\n",
       "         [ 0.4094, -0.0861, -0.7935,  ..., -0.1285, -0.9932, -0.3013],\n",
       "         ...,\n",
       "         [ 0.2688, -0.5630, -1.1902,  ...,  0.4493,  0.5404, -0.0103],\n",
       "         [ 0.0535,  0.4411,  0.5313,  ...,  0.0148, -1.0056,  0.3759],\n",
       "         [ 0.3031, -0.1590, -0.1316,  ..., -0.5384, -0.4271, -0.4876]],\n",
       "\n",
       "        [[-1.1865, -0.7280, -1.2331,  ..., -0.9013, -0.0542, -1.5949],\n",
       "         [-0.6345, -0.5920,  0.5326,  ..., -1.0395, -0.7963, -0.0647],\n",
       "         [-0.1132,  0.5166,  0.2569,  ...,  0.5595, -1.6881,  0.9485],\n",
       "         ...,\n",
       "         [-0.0254, -0.2669,  0.1927,  ..., -0.2917,  0.1088, -0.4807],\n",
       "         [-0.2609, -0.2328,  0.1404,  ..., -0.1325, -0.8436, -0.7524],\n",
       "         [-1.1399, -0.1751, -0.8705,  ...,  0.1589,  0.3377,  0.3493]]],\n",
       "       grad_fn=<SqueezeBackward1>)"
      ]
     },
     "execution_count": 79,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "torch.manual_seed(42)\n",
    "\n",
    "# Create a convolutional layer with same dimensions as TinyVGG \n",
    "# (try changing any of the parameters and see what happens)\n",
    "conv_layer = nn.Conv2d(in_channels=3,\n",
    "                       out_channels=10,\n",
    "                       kernel_size=3,\n",
    "                       stride=1,\n",
    "                       padding=0) # also try using \"valid\" or \"same\" here \n",
    "\n",
    "# Pass the data through the convolutional layer\n",
    "conv_layer(test_image) # Note: If running PyTorch <1.11.0, this will error because of shape issues (nn.Conv.2d() expects a 4d tensor as input) "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "cb0184ad-5c16-4e1c-bcfa-70ecf15377da",
   "metadata": {
    "id": "cb0184ad-5c16-4e1c-bcfa-70ecf15377da"
   },
   "source": [
    "If we try to pass a single image in, we get a shape mismatch error:\n",
    "\n",
    "> `RuntimeError: Expected 4-dimensional input for 4-dimensional weight [10, 3, 3, 3], but got 3-dimensional input of size [3, 64, 64] instead`\n",
    ">\n",
    "> **Note:** If you're running PyTorch 1.11.0+, this error won't occur.\n",
    "\n",
    "This is because our `nn.Conv2d()` layer expects a 4-dimensional tensor as input with size `(N, C, H, W)` or `[batch_size, color_channels, height, width]`.\n",
    "\n",
    "Right now our single image `test_image` only has a shape of `[color_channels, height, width]` or `[3, 64, 64]`.\n",
    "\n",
    "We can fix this for a single image using `test_image.unsqueeze(dim=0)` to add an extra dimension for `N`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "id": "abba741d-a1ed-44ed-ba53-41d589433a2c",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "abba741d-a1ed-44ed-ba53-41d589433a2c",
    "outputId": "9dd8151d-376c-4342-c379-91fcb6468706"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "torch.Size([1, 3, 64, 64])"
      ]
     },
     "execution_count": 80,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Add extra dimension to test image\n",
    "test_image.unsqueeze(dim=0).shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "id": "c7280a49-4ee0-452b-a514-61115b6a444c",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "c7280a49-4ee0-452b-a514-61115b6a444c",
    "outputId": "87bf7e37-c1a7-44a4-eef2-a02eb0489147"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "torch.Size([1, 10, 62, 62])"
      ]
     },
     "execution_count": 81,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Pass test image with extra dimension through conv_layer\n",
    "conv_layer(test_image.unsqueeze(dim=0)).shape"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "181df81b-7c5a-46cc-b8d5-a592bf755a13",
   "metadata": {
    "id": "181df81b-7c5a-46cc-b8d5-a592bf755a13"
   },
   "source": [
    "Hmm, notice what happens to our shape (the same shape as the first layer of TinyVGG on [CNN Explainer](https://poloclub.github.io/cnn-explainer/)), we get different channel sizes as well as different pixel sizes.\n",
    "\n",
    "What if we changed the values of `conv_layer`?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 101,
   "id": "04445d45-cf2f-4c1d-b215-bc50865a207a",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "04445d45-cf2f-4c1d-b215-bc50865a207a",
    "outputId": "eaa97fb8-52c0-493d-eac3-f2e23df9b01c"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "torch.Size([1, 10, 30, 30])"
      ]
     },
     "execution_count": 101,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "torch.manual_seed(42)\n",
    "# Create a new conv_layer with different values (try setting these to whatever you like)\n",
    "conv_layer_2 = nn.Conv2d(in_channels=3, # same number of color channels as our input image\n",
    "                         out_channels=10,\n",
    "                         kernel_size=(5, 5), # kernel is usually a square so a tuple also works\n",
    "                         stride=2,\n",
    "                         padding=0)\n",
    "\n",
    "# Pass single image through new conv_layer_2 (this calls nn.Conv2d()'s forward() method on the input)\n",
    "conv_layer_2(test_image.unsqueeze(dim=0)).shape"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b27dbdbb-3e32-4ffa-803e-cf943d96c72b",
   "metadata": {
    "id": "b27dbdbb-3e32-4ffa-803e-cf943d96c72b"
   },
   "source": [
    "Woah, we get another shape change.\n",
    "\n",
    "Now our image is of shape `[1, 10, 30, 30]` (it will be different if you use different values) or `[batch_size=1, color_channels=10, height=30, width=30]`.\n",
    "\n",
    "What's going on here?\n",
    "\n",
    "Behind the scenes, our `nn.Conv2d()` is compressing the information stored in the image.\n",
    "\n",
    "It does this by performing operations on the input (our test image) against its internal parameters.\n",
    "\n",
    "The goal of this is similar to all of the other neural networks we've been building.\n",
    "\n",
    "Data goes in and the layers try to update their internal parameters (patterns) to lower the loss function thanks to some help of the optimizer.\n",
    "\n",
    "The only difference is *how* the different layers calculate their parameter updates or in PyTorch terms, the operation present in the layer `forward()` method.\n",
    "\n",
    "If we check out our `conv_layer_2.state_dict()` we'll find a similar weight and bias setup as we've seen before."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "id": "46027ed1-c3a7-46bd-bab7-17f8c20e354b",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "46027ed1-c3a7-46bd-bab7-17f8c20e354b",
    "outputId": "bc493b18-1ef3-41c6-9de9-5dfeaa4c259e"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "OrderedDict([('weight', tensor([[[[ 0.0883,  0.0958, -0.0271,  0.1061, -0.0253],\n",
      "          [ 0.0233, -0.0562,  0.0678,  0.1018, -0.0847],\n",
      "          [ 0.1004,  0.0216,  0.0853,  0.0156,  0.0557],\n",
      "          [-0.0163,  0.0890,  0.0171, -0.0539,  0.0294],\n",
      "          [-0.0532, -0.0135, -0.0469,  0.0766, -0.0911]],\n",
      "\n",
      "         [[-0.0532, -0.0326, -0.0694,  0.0109, -0.1140],\n",
      "          [ 0.1043, -0.0981,  0.0891,  0.0192, -0.0375],\n",
      "          [ 0.0714,  0.0180,  0.0933,  0.0126, -0.0364],\n",
      "          [ 0.0310, -0.0313,  0.0486,  0.1031,  0.0667],\n",
      "          [-0.0505,  0.0667,  0.0207,  0.0586, -0.0704]],\n",
      "\n",
      "         [[-0.1143, -0.0446, -0.0886,  0.0947,  0.0333],\n",
      "          [ 0.0478,  0.0365, -0.0020,  0.0904, -0.0820],\n",
      "          [ 0.0073, -0.0788,  0.0356, -0.0398,  0.0354],\n",
      "          [-0.0241,  0.0958, -0.0684, -0.0689, -0.0689],\n",
      "          [ 0.1039,  0.0385,  0.1111, -0.0953, -0.1145]]],\n",
      "\n",
      "\n",
      "        [[[-0.0903, -0.0777,  0.0468,  0.0413,  0.0959],\n",
      "          [-0.0596, -0.0787,  0.0613, -0.0467,  0.0701],\n",
      "          [-0.0274,  0.0661, -0.0897, -0.0583,  0.0352],\n",
      "          [ 0.0244, -0.0294,  0.0688,  0.0785, -0.0837],\n",
      "          [-0.0616,  0.1057, -0.0390, -0.0409, -0.1117]],\n",
      "\n",
      "         [[-0.0661,  0.0288, -0.0152, -0.0838,  0.0027],\n",
      "          [-0.0789, -0.0980, -0.0636, -0.1011, -0.0735],\n",
      "          [ 0.1154,  0.0218,  0.0356, -0.1077, -0.0758],\n",
      "          [-0.0384,  0.0181, -0.1016, -0.0498, -0.0691],\n",
      "          [ 0.0003, -0.0430, -0.0080, -0.0782, -0.0793]],\n",
      "\n",
      "         [[-0.0674, -0.0395, -0.0911,  0.0968, -0.0229],\n",
      "          [ 0.0994,  0.0360, -0.0978,  0.0799, -0.0318],\n",
      "          [-0.0443, -0.0958, -0.1148,  0.0330, -0.0252],\n",
      "          [ 0.0450, -0.0948,  0.0857, -0.0848, -0.0199],\n",
      "          [ 0.0241,  0.0596,  0.0932,  0.1052, -0.0916]]],\n",
      "\n",
      "\n",
      "        [[[ 0.0291, -0.0497, -0.0127, -0.0864,  0.1052],\n",
      "          [-0.0847,  0.0617,  0.0406,  0.0375, -0.0624],\n",
      "          [ 0.1050,  0.0254,  0.0149, -0.1018,  0.0485],\n",
      "          [-0.0173, -0.0529,  0.0992,  0.0257, -0.0639],\n",
      "          [-0.0584, -0.0055,  0.0645, -0.0295, -0.0659]],\n",
      "\n",
      "         [[-0.0395, -0.0863,  0.0412,  0.0894, -0.1087],\n",
      "          [ 0.0268,  0.0597,  0.0209, -0.0411,  0.0603],\n",
      "          [ 0.0607,  0.0432, -0.0203, -0.0306,  0.0124],\n",
      "          [-0.0204, -0.0344,  0.0738,  0.0992, -0.0114],\n",
      "          [-0.0259,  0.0017, -0.0069,  0.0278,  0.0324]],\n",
      "\n",
      "         [[-0.1049, -0.0426,  0.0972,  0.0450, -0.0057],\n",
      "          [-0.0696, -0.0706, -0.1034, -0.0376,  0.0390],\n",
      "          [ 0.0736,  0.0533, -0.1021, -0.0694, -0.0182],\n",
      "          [ 0.1117,  0.0167, -0.0299,  0.0478, -0.0440],\n",
      "          [-0.0747,  0.0843, -0.0525, -0.0231, -0.1149]]],\n",
      "\n",
      "\n",
      "        [[[ 0.0773,  0.0875,  0.0421, -0.0805, -0.1140],\n",
      "          [-0.0938,  0.0861,  0.0554,  0.0972,  0.0605],\n",
      "          [ 0.0292, -0.0011, -0.0878, -0.0989, -0.1080],\n",
      "          [ 0.0473, -0.0567, -0.0232, -0.0665, -0.0210],\n",
      "          [-0.0813, -0.0754,  0.0383, -0.0343,  0.0713]],\n",
      "\n",
      "         [[-0.0370, -0.0847, -0.0204, -0.0560, -0.0353],\n",
      "          [-0.1099,  0.0646, -0.0804,  0.0580,  0.0524],\n",
      "          [ 0.0825, -0.0886,  0.0830, -0.0546,  0.0428],\n",
      "          [ 0.1084, -0.0163, -0.0009, -0.0266, -0.0964],\n",
      "          [ 0.0554, -0.1146,  0.0717,  0.0864,  0.1092]],\n",
      "\n",
      "         [[-0.0272, -0.0949,  0.0260,  0.0638, -0.1149],\n",
      "          [-0.0262, -0.0692, -0.0101, -0.0568, -0.0472],\n",
      "          [-0.0367, -0.1097,  0.0947,  0.0968, -0.0181],\n",
      "          [-0.0131, -0.0471, -0.1043, -0.1124,  0.0429],\n",
      "          [-0.0634, -0.0742, -0.0090, -0.0385, -0.0374]]],\n",
      "\n",
      "\n",
      "        [[[ 0.0037, -0.0245, -0.0398, -0.0553, -0.0940],\n",
      "          [ 0.0968, -0.0462,  0.0306, -0.0401,  0.0094],\n",
      "          [ 0.1077,  0.0532, -0.1001,  0.0458,  0.1096],\n",
      "          [ 0.0304,  0.0774,  0.1138, -0.0177,  0.0240],\n",
      "          [-0.0803, -0.0238,  0.0855,  0.0592, -0.0731]],\n",
      "\n",
      "         [[-0.0926, -0.0789, -0.1140, -0.0891, -0.0286],\n",
      "          [ 0.0779,  0.0193, -0.0878, -0.0926,  0.0574],\n",
      "          [-0.0859, -0.0142,  0.0554, -0.0534, -0.0126],\n",
      "          [-0.0101, -0.0273, -0.0585, -0.1029, -0.0933],\n",
      "          [-0.0618,  0.1115, -0.0558, -0.0775,  0.0280]],\n",
      "\n",
      "         [[ 0.0318,  0.0633,  0.0878,  0.0643, -0.1145],\n",
      "          [ 0.0102,  0.0699, -0.0107, -0.0680,  0.1101],\n",
      "          [-0.0432, -0.0657, -0.1041,  0.0052,  0.0512],\n",
      "          [ 0.0256,  0.0228, -0.0876, -0.1078,  0.0020],\n",
      "          [ 0.1053,  0.0666, -0.0672, -0.0150, -0.0851]]],\n",
      "\n",
      "\n",
      "        [[[-0.0557,  0.0209,  0.0629,  0.0957, -0.1060],\n",
      "          [ 0.0772, -0.0814,  0.0432,  0.0977,  0.0016],\n",
      "          [ 0.1051, -0.0984, -0.0441,  0.0673, -0.0252],\n",
      "          [-0.0236, -0.0481,  0.0796,  0.0566,  0.0370],\n",
      "          [-0.0649, -0.0937,  0.0125,  0.0342, -0.0533]],\n",
      "\n",
      "         [[-0.0323,  0.0780,  0.0092,  0.0052, -0.0284],\n",
      "          [-0.1046, -0.1086, -0.0552, -0.0587,  0.0360],\n",
      "          [-0.0336, -0.0452,  0.1101,  0.0402,  0.0823],\n",
      "          [-0.0559, -0.0472,  0.0424, -0.0769, -0.0755],\n",
      "          [-0.0056, -0.0422, -0.0866,  0.0685,  0.0929]],\n",
      "\n",
      "         [[ 0.0187, -0.0201, -0.1070, -0.0421,  0.0294],\n",
      "          [ 0.0544, -0.0146, -0.0457,  0.0643, -0.0920],\n",
      "          [ 0.0730, -0.0448,  0.0018, -0.0228,  0.0140],\n",
      "          [-0.0349,  0.0840, -0.0030,  0.0901,  0.1110],\n",
      "          [-0.0563, -0.0842,  0.0926,  0.0905, -0.0882]]],\n",
      "\n",
      "\n",
      "        [[[-0.0089, -0.1139, -0.0945,  0.0223,  0.0307],\n",
      "          [ 0.0245, -0.0314,  0.1065,  0.0165, -0.0681],\n",
      "          [-0.0065,  0.0277,  0.0404, -0.0816,  0.0433],\n",
      "          [-0.0590, -0.0959, -0.0631,  0.1114,  0.0987],\n",
      "          [ 0.1034,  0.0678,  0.0872, -0.0155, -0.0635]],\n",
      "\n",
      "         [[ 0.0577, -0.0598, -0.0779, -0.0369,  0.0242],\n",
      "          [ 0.0594, -0.0448, -0.0680,  0.0156, -0.0681],\n",
      "          [-0.0752,  0.0602, -0.0194,  0.1055,  0.1123],\n",
      "          [ 0.0345,  0.0397,  0.0266,  0.0018, -0.0084],\n",
      "          [ 0.0016,  0.0431,  0.1074, -0.0299, -0.0488]],\n",
      "\n",
      "         [[-0.0280, -0.0558,  0.0196,  0.0862,  0.0903],\n",
      "          [ 0.0530, -0.0850, -0.0620, -0.0254, -0.0213],\n",
      "          [ 0.0095, -0.1060,  0.0359, -0.0881, -0.0731],\n",
      "          [-0.0960,  0.1006, -0.1093,  0.0871, -0.0039],\n",
      "          [-0.0134,  0.0722, -0.0107,  0.0724,  0.0835]]],\n",
      "\n",
      "\n",
      "        [[[-0.1003,  0.0444,  0.0218,  0.0248,  0.0169],\n",
      "          [ 0.0316, -0.0555, -0.0148,  0.1097,  0.0776],\n",
      "          [-0.0043, -0.1086,  0.0051, -0.0786,  0.0939],\n",
      "          [-0.0701, -0.0083, -0.0256,  0.0205,  0.1087],\n",
      "          [ 0.0110,  0.0669,  0.0896,  0.0932, -0.0399]],\n",
      "\n",
      "         [[-0.0258,  0.0556, -0.0315,  0.0541, -0.0252],\n",
      "          [-0.0783,  0.0470,  0.0177,  0.0515,  0.1147],\n",
      "          [ 0.0788,  0.1095,  0.0062, -0.0993, -0.0810],\n",
      "          [-0.0717, -0.1018, -0.0579, -0.1063, -0.1065],\n",
      "          [-0.0690, -0.1138, -0.0709,  0.0440,  0.0963]],\n",
      "\n",
      "         [[-0.0343, -0.0336,  0.0617, -0.0570, -0.0546],\n",
      "          [ 0.0711, -0.1006,  0.0141,  0.1020,  0.0198],\n",
      "          [ 0.0314, -0.0672, -0.0016,  0.0063,  0.0283],\n",
      "          [ 0.0449,  0.1003, -0.0881,  0.0035, -0.0577],\n",
      "          [-0.0913, -0.0092, -0.1016,  0.0806,  0.0134]]],\n",
      "\n",
      "\n",
      "        [[[-0.0622,  0.0603, -0.1093, -0.0447, -0.0225],\n",
      "          [-0.0981, -0.0734, -0.0188,  0.0876,  0.1115],\n",
      "          [ 0.0735, -0.0689, -0.0755,  0.1008,  0.0408],\n",
      "          [ 0.0031,  0.0156, -0.0928, -0.0386,  0.1112],\n",
      "          [-0.0285, -0.0058, -0.0959, -0.0646, -0.0024]],\n",
      "\n",
      "         [[-0.0717, -0.0143,  0.0470, -0.1130,  0.0343],\n",
      "          [-0.0763, -0.0564,  0.0443,  0.0918, -0.0316],\n",
      "          [-0.0474, -0.1044, -0.0595, -0.1011, -0.0264],\n",
      "          [ 0.0236, -0.1082,  0.1008,  0.0724, -0.1130],\n",
      "          [-0.0552,  0.0377, -0.0237, -0.0126, -0.0521]],\n",
      "\n",
      "         [[ 0.0927, -0.0645,  0.0958,  0.0075,  0.0232],\n",
      "          [ 0.0901, -0.0190, -0.0657, -0.0187,  0.0937],\n",
      "          [-0.0857,  0.0262, -0.1135,  0.0605,  0.0427],\n",
      "          [ 0.0049,  0.0496,  0.0001,  0.0639, -0.0914],\n",
      "          [-0.0170,  0.0512,  0.1150,  0.0588, -0.0840]]],\n",
      "\n",
      "\n",
      "        [[[ 0.0888, -0.0257, -0.0247, -0.1050, -0.0182],\n",
      "          [ 0.0817,  0.0161, -0.0673,  0.0355, -0.0370],\n",
      "          [ 0.1054, -0.1002, -0.0365, -0.1115, -0.0455],\n",
      "          [ 0.0364,  0.1112,  0.0194,  0.1132,  0.0226],\n",
      "          [ 0.0667,  0.0926,  0.0965, -0.0646,  0.1062]],\n",
      "\n",
      "         [[ 0.0699, -0.0540, -0.0551, -0.0969,  0.0290],\n",
      "          [-0.0936,  0.0488,  0.0365, -0.1003,  0.0315],\n",
      "          [-0.0094,  0.0527,  0.0663, -0.1148,  0.1059],\n",
      "          [ 0.0968,  0.0459, -0.1055, -0.0412, -0.0335],\n",
      "          [-0.0297,  0.0651,  0.0420,  0.0915, -0.0432]],\n",
      "\n",
      "         [[ 0.0389,  0.0411, -0.0961, -0.1120, -0.0599],\n",
      "          [ 0.0790, -0.1087, -0.1005,  0.0647,  0.0623],\n",
      "          [ 0.0950, -0.0872, -0.0845,  0.0592,  0.1004],\n",
      "          [ 0.0691,  0.0181,  0.0381,  0.1096, -0.0745],\n",
      "          [-0.0524,  0.0808, -0.0790, -0.0637,  0.0843]]]])), ('bias', tensor([ 0.0364,  0.0373, -0.0489, -0.0016,  0.1057, -0.0693,  0.0009,  0.0549,\n",
      "        -0.0797,  0.1121]))])\n"
     ]
    }
   ],
   "source": [
    "# Check out the conv_layer_2 internal parameters\n",
    "print(conv_layer_2.state_dict())"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8b708eb6-ae46-4d8c-a8a4-1392827d3e37",
   "metadata": {
    "id": "8b708eb6-ae46-4d8c-a8a4-1392827d3e37"
   },
   "source": [
    "Look at that! A bunch of random numbers for a weight and bias tensor.\n",
    "\n",
    "The shapes of these are manipulated by the inputs we passed to `nn.Conv2d()` when we set it up.\n",
    "\n",
    "Let's check them out."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "id": "e5518d61-c0b7-4351-b5ea-4d6b6144291a",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "e5518d61-c0b7-4351-b5ea-4d6b6144291a",
    "outputId": "14ef701e-c82d-4fae-9dd6-e18352817cf2"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "conv_layer_2 weight shape: \n",
      "torch.Size([10, 3, 5, 5]) -> [out_channels=10, in_channels=3, kernel_size=5, kernel_size=5]\n",
      "\n",
      "conv_layer_2 bias shape: \n",
      "torch.Size([10]) -> [out_channels=10]\n"
     ]
    }
   ],
   "source": [
    "# Get shapes of weight and bias tensors within conv_layer_2\n",
    "print(f\"conv_layer_2 weight shape: \\n{conv_layer_2.weight.shape} -> [out_channels=10, in_channels=3, kernel_size=5, kernel_size=5]\")\n",
    "print(f\"\\nconv_layer_2 bias shape: \\n{conv_layer_2.bias.shape} -> [out_channels=10]\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f0de23c7-4501-4156-80a4-ac889a636a42",
   "metadata": {
    "id": "f0de23c7-4501-4156-80a4-ac889a636a42"
   },
   "source": [
    "> **Question:** What should we set the parameters of our `nn.Conv2d()` layers?\n",
    ">\n",
    "> That's a good one. But similar to many other things in machine learning, the values of these aren't set in stone (and recall, because these values are ones we can set ourselves, they're referred to as \"**hyperparameters**\"). \n",
    ">\n",
    "> The best way to find out is to try out different values and see how they effect your model's performance.\n",
    ">\n",
    "> Or even better, find a working example on a problem similar to yours (like we've done with TinyVGG) and copy it. \n",
    "\n",
    "We're working with a different of layer here to what we've seen before.\n",
    "\n",
    "But the premise remains the same: start with random numbers and update them to better represent the data."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6370d45d-ca44-4fa0-a2d7-efaf0a207b91",
   "metadata": {
    "id": "6370d45d-ca44-4fa0-a2d7-efaf0a207b91"
   },
   "source": [
    "### 7.2 Stepping through `nn.MaxPool2d()`\n",
    "Now let's check out what happens when we move data through `nn.MaxPool2d()`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "id": "1164c753-19d9-43b7-a04f-017d0f7188c3",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "1164c753-19d9-43b7-a04f-017d0f7188c3",
    "outputId": "9c46f08e-928d-4ee4-e43c-b402113fc2b4"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Test image original shape: torch.Size([3, 64, 64])\n",
      "Test image with unsqueezed dimension: torch.Size([1, 3, 64, 64])\n",
      "Shape after going through conv_layer(): torch.Size([1, 10, 62, 62])\n",
      "Shape after going through conv_layer() and max_pool_layer(): torch.Size([1, 10, 31, 31])\n"
     ]
    }
   ],
   "source": [
    "# Print out original image shape without and with unsqueezed dimension\n",
    "print(f\"Test image original shape: {test_image.shape}\")\n",
    "print(f\"Test image with unsqueezed dimension: {test_image.unsqueeze(dim=0).shape}\")\n",
    "\n",
    "# Create a sample nn.MaxPoo2d() layer\n",
    "max_pool_layer = nn.MaxPool2d(kernel_size=2)\n",
    "\n",
    "# Pass data through just the conv_layer\n",
    "test_image_through_conv = conv_layer(test_image.unsqueeze(dim=0))\n",
    "print(f\"Shape after going through conv_layer(): {test_image_through_conv.shape}\")\n",
    "\n",
    "# Pass data through the max pool layer\n",
    "test_image_through_conv_and_max_pool = max_pool_layer(test_image_through_conv)\n",
    "print(f\"Shape after going through conv_layer() and max_pool_layer(): {test_image_through_conv_and_max_pool.shape}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "de029abd-6674-4bfa-99ab-322f339f89f4",
   "metadata": {
    "id": "de029abd-6674-4bfa-99ab-322f339f89f4"
   },
   "source": [
    "Notice the change in the shapes of what's happening in and out of a `nn.MaxPool2d()` layer.\n",
    "\n",
    "The `kernel_size` of the `nn.MaxPool2d()` layer will effects the size of the output shape.\n",
    "\n",
    "In our case, the shape halves from a `62x62` image to `31x31` image.\n",
    "\n",
    "Let's see this work with a smaller tensor."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "id": "e6a2b196-4845-4b40-9212-e75406e88875",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "e6a2b196-4845-4b40-9212-e75406e88875",
    "outputId": "5a5e5df1-8e25-4061-d223-398ecfc7ef4c"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Random tensor:\n",
      "tensor([[[[0.3367, 0.1288],\n",
      "          [0.2345, 0.2303]]]])\n",
      "Random tensor shape: torch.Size([1, 1, 2, 2])\n",
      "\n",
      "Max pool tensor:\n",
      "tensor([[[[0.3367]]]]) <- this is the maximum value from random_tensor\n",
      "Max pool tensor shape: torch.Size([1, 1, 1, 1])\n"
     ]
    }
   ],
   "source": [
    "torch.manual_seed(42)\n",
    "# Create a random tensor with a similiar number of dimensions to our images\n",
    "random_tensor = torch.randn(size=(1, 1, 2, 2))\n",
    "print(f\"Random tensor:\\n{random_tensor}\")\n",
    "print(f\"Random tensor shape: {random_tensor.shape}\")\n",
    "\n",
    "# Create a max pool layer\n",
    "max_pool_layer = nn.MaxPool2d(kernel_size=2) # see what happens when you change the kernel_size value \n",
    "\n",
    "# Pass the random tensor through the max pool layer\n",
    "max_pool_tensor = max_pool_layer(random_tensor)\n",
    "print(f\"\\nMax pool tensor:\\n{max_pool_tensor} <- this is the maximum value from random_tensor\")\n",
    "print(f\"Max pool tensor shape: {max_pool_tensor.shape}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "002e586e-dcb3-40fe-a7dd-a1c18a3b8da0",
   "metadata": {
    "id": "002e586e-dcb3-40fe-a7dd-a1c18a3b8da0"
   },
   "source": [
    "Notice the final two dimensions between `random_tensor` and `max_pool_tensor`, they go from `[2, 2]` to `[1, 1]`.\n",
    "\n",
    "In essence, they get halved.\n",
    "\n",
    "And the change would be different for different values of `kernel_size` for `nn.MaxPool2d()`.\n",
    "\n",
    "Also notice the value leftover in `max_pool_tensor` is the **maximum** value from `random_tensor`.\n",
    "\n",
    "What's happening here?\n",
    "\n",
    "This is another important piece of the puzzle of neural networks.\n",
    "\n",
    "Essentially, **every layer in a neural network is trying to compress data from higher dimensional space to lower dimensional space**. \n",
    "\n",
    "In other words, take a lot of numbers (raw data) and learn patterns in those numbers, patterns that are predictive whilst also being *smaller* in size than the original values.\n",
    "\n",
    "From an artificial intelligence perspective, you could consider the whole goal of a neural network to *compress* information.\n",
    "\n",
    "![each layer of a neural network compresses the original input data into a smaller representation that is (hopefully) capable of making predictions on future input data](https://raw.githubusercontent.com/mrdbourke/pytorch-deep-learning/main/images/03-conv-net-as-compression.png)\n",
    "\n",
    "This means, that from the point of view of a neural network, intelligence is compression.\n",
    "\n",
    "This is the idea of the use of a `nn.MaxPool2d()` layer: take the maximum value from a portion of a tensor and disregard the rest.\n",
    "\n",
    "In essence, lowering the dimensionality of a tensor whilst still retaining a (hopefully) significant portion of the information.\n",
    "\n",
    "It is the same story for a `nn.Conv2d()` layer.\n",
    "\n",
    "Except instead of just taking the maximum, the `nn.Conv2d()` performs a convolutional operation on the data (see this in action on the [CNN Explainer webpage](https://poloclub.github.io/cnn-explainer/)).\n",
    "\n",
    "> **Exercise:** What do you think the [`nn.AvgPool2d()`](https://pytorch.org/docs/stable/generated/torch.nn.AvgPool2d.html) layer does? Try making a random tensor like we did above and passing it through. Check the input and output shapes as well as the input and output values.\n",
    "\n",
    "> **Extra-curriculum:** Lookup \"most common convolutional neural networks\", what architectures do you find? Are any of them contained within the [`torchvision.models`](https://pytorch.org/vision/stable/models.html) library? What do you think you could do with these?"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "39a3c646-52f0-4f4b-8527-2fc33d0dfb13",
   "metadata": {
    "id": "39a3c646-52f0-4f4b-8527-2fc33d0dfb13"
   },
   "source": [
    "### 7.3 Setup a loss function and optimizer for `model_2`\n",
    "\n",
    "We've stepped through the layers in our first CNN enough.\n",
    "\n",
    "But remember, if something still isn't clear, try starting small.\n",
    "\n",
    "Pick a single layer of a model, pass some data through it and see what happens.\n",
    "\n",
    "Now it's time to move forward and get to training!\n",
    "\n",
    "Let's setup a loss function and an optimizer.\n",
    "\n",
    "We'll use the functions as before, `nn.CrossEntropyLoss()` as the loss function (since we're working with multi-class classification data).\n",
    "\n",
    "And `torch.optim.SGD()` as the optimizer to optimize `model_2.parameters()` with a learning rate of `0.1`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 88,
   "id": "06a76a1b-5f6f-4018-bf7b-8388b385476f",
   "metadata": {
    "id": "06a76a1b-5f6f-4018-bf7b-8388b385476f"
   },
   "outputs": [],
   "source": [
    "# Setup loss and optimizer\n",
    "loss_fn = nn.CrossEntropyLoss()\n",
    "optimizer = torch.optim.SGD(params=model_2.parameters(), \n",
    "                             lr=0.1)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "758bc223-a244-4604-a07a-e2fc2f96c2f6",
   "metadata": {
    "id": "758bc223-a244-4604-a07a-e2fc2f96c2f6"
   },
   "source": [
    "### 7.4 Training and testing `model_2` using our training and test functions\n",
    "\n",
    "Loss and optimizer ready!\n",
    "\n",
    "Time to train and test.\n",
    "\n",
    "We'll use our `train_step()` and `test_step()` functions we created before.\n",
    "\n",
    "We'll also measure the time to compare it to our other models."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 89,
   "id": "861d126e-d876-40b3-9b7a-66cfc2f1bf05",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 327,
     "referenced_widgets": [
      "2b9c90ceb8554eaaaaf33acacecbcc11",
      "9051de473b1b456592576115260f0c48",
      "a315b7b535e4461ca11a6d96ca74411c",
      "4cd01ed6c2534d4e80a0af2e9da02052",
      "eb3ca30526e24fff9194d4e82436df99",
      "6117713ba6c9490fab837cfbad7d442b",
      "d426a4e9fff447ea95c7256372e272b7",
      "fdcf4db7208d42b5acb35bf56c72fd82",
      "606df8221adf48308f48c86ac54475e8",
      "ebc6218d07f54005b0e46553b83464e1",
      "03435ac81db84ae6bb6cb99bb78167c2"
     ]
    },
    "id": "861d126e-d876-40b3-9b7a-66cfc2f1bf05",
    "outputId": "77ae1601-bf72-4239-b3f6-9fb99a1c62c0"
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "  0%|          | 0/3 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch: 0\n",
      "---------\n",
      "Train loss: 0.53990 | Train accuracy: 80.71%\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      " 33%|███▎      | 1/3 [00:22<00:45, 22.93s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Test loss: 0.37613 | Test accuracy: 86.66%\n",
      "\n",
      "Epoch: 1\n",
      "---------\n",
      "Train loss: 0.34169 | Train accuracy: 87.91%\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      " 67%|██████▋   | 2/3 [00:45<00:22, 22.43s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Test loss: 0.33961 | Test accuracy: 87.74%\n",
      "\n",
      "Epoch: 2\n",
      "---------\n",
      "Train loss: 0.30690 | Train accuracy: 89.13%\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 3/3 [01:07<00:00, 22.37s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Test loss: 0.30596 | Test accuracy: 89.17%\n",
      "\n",
      "Train time on cuda: 67.100 seconds\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "torch.manual_seed(42)\n",
    "torch.cuda.manual_seed(42)\n",
    "\n",
    "# Measure time\n",
    "from timeit import default_timer as timer\n",
    "train_time_start_model_2 = timer()\n",
    "\n",
    "# Train and test model \n",
    "epochs = 3\n",
    "for epoch in tqdm(range(epochs)):\n",
    "    print(f\"Epoch: {epoch}\\n---------\")\n",
    "    train_step(data_loader=train_dataloader, \n",
    "        model=model_2, \n",
    "        loss_fn=loss_fn,\n",
    "        optimizer=optimizer,\n",
    "        accuracy_fn=accuracy_fn,\n",
    "        device=device\n",
    "    )\n",
    "    test_step(data_loader=test_dataloader,\n",
    "        model=model_2,\n",
    "        loss_fn=loss_fn,\n",
    "        accuracy_fn=accuracy_fn,\n",
    "        device=device\n",
    "    )\n",
    "\n",
    "train_time_end_model_2 = timer()\n",
    "total_train_time_model_2 = print_train_time(start=train_time_start_model_2,\n",
    "                                           end=train_time_end_model_2,\n",
    "                                           device=device)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "cfec7b7e-4dba-4016-957a-a29c6c10fde0",
   "metadata": {
    "id": "cfec7b7e-4dba-4016-957a-a29c6c10fde0"
   },
   "source": [
    "Woah! Looks like the convolutional and max pooling layers helped improve performance a little.\n",
    "\n",
    "Let's evaluate `model_2`'s results with our `eval_model()` function."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 90,
   "id": "c1bf8b89-1389-4395-a1c4-9c6e94d9e71c",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "c1bf8b89-1389-4395-a1c4-9c6e94d9e71c",
    "outputId": "6ccddfc5-a86b-409d-d329-0d140f5b2f10"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'model_name': 'FashionMNISTModelV2',\n",
       " 'model_loss': 0.30595821142196655,\n",
       " 'model_acc': 89.1673322683706}"
      ]
     },
     "execution_count": 90,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Get model_2 results \n",
    "model_2_results = eval_model(\n",
    "    model=model_2,\n",
    "    data_loader=test_dataloader,\n",
    "    loss_fn=loss_fn,\n",
    "    accuracy_fn=accuracy_fn\n",
    ")\n",
    "model_2_results"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "24c5ff68-b0bc-4b09-9da6-696736bc262d",
   "metadata": {
    "id": "24c5ff68-b0bc-4b09-9da6-696736bc262d"
   },
   "source": [
    "## 8. Compare model results and training time\n",
    "\n",
    "We've trained three different models.\n",
    "\n",
    "1. `model_0` - our baseline model with two `nn.Linear()` layers.\n",
    "2. `model_1` - the same setup as our baseline model except with `nn.ReLU()` layers in between the `nn.Linear()` layers.\n",
    "3. `model_2` - our first CNN model that mimics the TinyVGG architecture on the CNN Explainer website.\n",
    "\n",
    "This is a regular practice in machine learning.\n",
    "\n",
    "Building multiple models and performing multiple training experiments to see which performs best.\n",
    "\n",
    "Let's combine our model results dictionaries into a DataFrame and find out."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 100,
   "id": "52d84ee1-1ad4-4860-b147-f8912c1febc7",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 143
    },
    "id": "52d84ee1-1ad4-4860-b147-f8912c1febc7",
    "outputId": "6eebb1ea-293a-4ae1-c054-f6154868426c"
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>model_name</th>\n",
       "      <th>model_loss</th>\n",
       "      <th>model_acc</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>FashionMNISTModelV0</td>\n",
       "      <td>0.501004</td>\n",
       "      <td>83.126997</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>FashionMNISTModelV1</td>\n",
       "      <td>0.460156</td>\n",
       "      <td>84.195288</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>FashionMNISTModelV2</td>\n",
       "      <td>0.305958</td>\n",
       "      <td>89.167332</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            model_name  model_loss  model_acc\n",
       "0  FashionMNISTModelV0    0.501004  83.126997\n",
       "1  FashionMNISTModelV1    0.460156  84.195288\n",
       "2  FashionMNISTModelV2    0.305958  89.167332"
      ]
     },
     "execution_count": 100,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import pandas as pd\n",
    "compare_results = pd.DataFrame([model_0_results, model_1_results, model_2_results])\n",
    "compare_results"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c67f3fb5-ce7b-40b8-86a0-2797492de0ef",
   "metadata": {
    "id": "c67f3fb5-ce7b-40b8-86a0-2797492de0ef"
   },
   "source": [
    "Nice!\n",
    "\n",
    "We can add the training time values too."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 102,
   "id": "297af38f-e69f-4c6f-9027-fcaf0482a55c",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 143
    },
    "id": "297af38f-e69f-4c6f-9027-fcaf0482a55c",
    "outputId": "67c01781-78c2-47e8-a1ee-452869bea5e2"
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>model_name</th>\n",
       "      <th>model_loss</th>\n",
       "      <th>model_acc</th>\n",
       "      <th>training_time</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>FashionMNISTModelV0</td>\n",
       "      <td>0.501004</td>\n",
       "      <td>83.126997</td>\n",
       "      <td>54.386199</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>FashionMNISTModelV1</td>\n",
       "      <td>0.460156</td>\n",
       "      <td>84.195288</td>\n",
       "      <td>58.957129</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>FashionMNISTModelV2</td>\n",
       "      <td>0.305958</td>\n",
       "      <td>89.167332</td>\n",
       "      <td>67.099776</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            model_name  model_loss  model_acc  training_time\n",
       "0  FashionMNISTModelV0    0.501004  83.126997      54.386199\n",
       "1  FashionMNISTModelV1    0.460156  84.195288      58.957129\n",
       "2  FashionMNISTModelV2    0.305958  89.167332      67.099776"
      ]
     },
     "execution_count": 102,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Add training times to results comparison\n",
    "compare_results[\"training_time\"] = [total_train_time_model_0,\n",
    "                                    total_train_time_model_1,\n",
    "                                    total_train_time_model_2]\n",
    "compare_results"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "fbbe5832-1081-4c76-8d5b-06c7a06da7b9",
   "metadata": {
    "id": "fbbe5832-1081-4c76-8d5b-06c7a06da7b9"
   },
   "source": [
    "It looks like our CNN (`FashionMNISTModelV2`) model performed the best (lowest loss, highest accuracy) but had the longest training time.\n",
    "\n",
    "And our baseline model (`FashionMNISTModelV0`) performed better than `model_1` (`FashionMNISTModelV1`).\n",
    "\n",
    "### Performance-speed tradeoff\n",
    "\n",
    "Something to be aware of in machine learning is the **performance-speed** tradeoff.\n",
    "\n",
    "Generally, you get better performance out of a larger, more complex model (like we did with `model_2`).\n",
    "\n",
    "However, this performance increase often comes at a sacrifice of training speed and inference speed.\n",
    "\n",
    "> **Note:** The training times you get will be very dependant on the hardware you use. \n",
    ">\n",
    "> Generally, the more CPU cores you have, the faster your models will train on CPU. And similar for GPUs.\n",
    "> \n",
    "> Newer hardware (in terms of age) will also often train models faster due to incorporating technology advances.\n",
    "\n",
    "How about we get visual?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 103,
   "id": "5eb0df60-9318-47d0-adce-f8788ed3999e",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 449
    },
    "id": "5eb0df60-9318-47d0-adce-f8788ed3999e",
    "outputId": "d904e039-1544-49f5-bed7-b17555d03b5a"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Visualize our model results\n",
    "compare_results.set_index(\"model_name\")[\"model_acc\"].plot(kind=\"barh\")\n",
    "plt.xlabel(\"accuracy (%)\")\n",
    "plt.ylabel(\"model\");"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 104,
   "id": "9e135a43",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Visualize our model results\n",
    "compare_results.set_index(\"model_name\")[\"training_time\"].plot(kind=\"barh\")\n",
    "plt.xlabel(\"training time [s]\")\n",
    "plt.ylabel(\"model\");"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0ba50d51-adb3-4e49-9b9a-85173e747352",
   "metadata": {
    "id": "0ba50d51-adb3-4e49-9b9a-85173e747352"
   },
   "source": [
    "## 9. Make and evaluate random predictions with best model\n",
    "\n",
    "Alright, we've compared our models to each other, let's further evaluate our best performing model, `model_2`.\n",
    "\n",
    "To do so, let's create a function `make_predictions()` where we can pass the model and some data for it to predict on."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 105,
   "id": "d1d5d3e7-9601-4141-8bd7-9abbd016bf6c",
   "metadata": {
    "id": "d1d5d3e7-9601-4141-8bd7-9abbd016bf6c"
   },
   "outputs": [],
   "source": [
    "def make_predictions(model: torch.nn.Module, data: list, device: torch.device = device):\n",
    "    pred_probs = []\n",
    "    model.eval()\n",
    "    with torch.inference_mode():\n",
    "        for sample in data:\n",
    "            # Prepare sample\n",
    "            sample = torch.unsqueeze(sample, dim=0).to(device) # Add an extra dimension and send sample to device\n",
    "\n",
    "            # Forward pass (model outputs raw logit)\n",
    "            pred_logit = model(sample)\n",
    "\n",
    "            # Get prediction probability (logit -> prediction probability)\n",
    "            pred_prob = torch.softmax(pred_logit.squeeze(), dim=0) # note: perform softmax on the \"logits\" dimension, not \"batch\" dimension (in this case we have a batch size of 1, so can perform on dim=0)\n",
    "\n",
    "            # Get pred_prob off GPU for further calculations\n",
    "            pred_probs.append(pred_prob.cpu())\n",
    "            \n",
    "    # Stack the pred_probs to turn list into a tensor\n",
    "    return torch.stack(pred_probs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 119,
   "id": "c8477701",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "9"
      ]
     },
     "execution_count": 119,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "list(test_data)[0][1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 106,
   "id": "36e3d6e7",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x1de21fd24d0>"
      ]
     },
     "execution_count": 106,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "img, label = test_data[1]\n",
    "plt.imshow(img.squeeze(), cmap='gray')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 107,
   "id": "420c7461-eaa9-4459-9e68-53574c758765",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "420c7461-eaa9-4459-9e68-53574c758765",
    "outputId": "f3dd6437-4f0f-4bc2-f9e6-d0969df63a52"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Test sample image shape: torch.Size([1, 28, 28])\n",
      "Test sample label: 5 (Sandal)\n"
     ]
    }
   ],
   "source": [
    "import random\n",
    "random.seed(42)\n",
    "N_SAMPLES = 16\n",
    "test_samples = []\n",
    "test_labels = []\n",
    "for sample, label in random.sample(list(test_data), k=N_SAMPLES):\n",
    "    test_samples.append(sample)\n",
    "    test_labels.append(label)\n",
    "\n",
    "# View the first test sample shape and label\n",
    "print(f\"Test sample image shape: {test_samples[0].shape}\\nTest sample label: {test_labels[0]} ({class_names[test_labels[0]]})\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 108,
   "id": "1DYqA0r4SkrV",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "1DYqA0r4SkrV",
    "outputId": "97bc573d-b39b-4eb2-caad-0379257b555e"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tensor([[1.7907e-07, 1.6592e-09, 9.7918e-09, 6.0261e-09, 2.5302e-08, 9.9990e-01,\n",
       "         4.0374e-07, 6.9657e-08, 2.7098e-07, 9.6975e-05],\n",
       "        [1.4180e-01, 7.3204e-01, 3.0735e-04, 1.2457e-02, 4.2268e-02, 3.8045e-05,\n",
       "         6.9792e-02, 6.9207e-05, 1.0074e-03, 2.1351e-04],\n",
       "        [1.5584e-04, 1.7424e-06, 1.1760e-05, 8.6596e-05, 8.9731e-05, 3.3562e-03,\n",
       "         2.9278e-05, 3.8030e-01, 3.2506e-03, 6.1272e-01],\n",
       "        [3.1837e-04, 6.5321e-05, 5.4241e-02, 2.8973e-04, 9.4209e-01, 1.0533e-05,\n",
       "         2.9301e-03, 6.9103e-07, 4.7505e-05, 8.6932e-06],\n",
       "        [1.1993e-02, 4.3525e-05, 4.0219e-05, 9.7806e-01, 1.1585e-03, 2.5164e-06,\n",
       "         8.4038e-03, 3.3815e-06, 2.8300e-04, 1.2492e-05],\n",
       "        [6.0800e-01, 4.4629e-05, 5.2030e-04, 6.9307e-03, 1.2024e-03, 1.0054e-06,\n",
       "         3.8266e-01, 6.5042e-07, 6.3331e-04, 2.9335e-07],\n",
       "        [5.0041e-05, 1.7613e-05, 3.7344e-03, 1.4327e-06, 9.8364e-01, 1.2429e-06,\n",
       "         1.2523e-02, 6.0527e-09, 2.4546e-05, 5.0132e-06],\n",
       "        [2.0975e-07, 1.3224e-08, 7.3293e-07, 3.2334e-07, 8.0971e-08, 2.8735e-03,\n",
       "         1.1265e-07, 9.9709e-01, 3.2039e-05, 5.6873e-06],\n",
       "        [2.0427e-07, 9.9999e-01, 5.8602e-08, 5.3570e-06, 1.3101e-07, 5.6713e-09,\n",
       "         8.8476e-10, 1.1492e-08, 7.1379e-09, 3.9025e-08]])"
      ]
     },
     "execution_count": 108,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Make predictions on test samples with model 2\n",
    "pred_probs= make_predictions(model=model_2, \n",
    "                             data=test_samples)\n",
    "\n",
    "# View first two prediction probabilities list\n",
    "pred_probs[:9]"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e9f40dd9-7987-42a9-84cc-65dc912a6345",
   "metadata": {
    "id": "e9f40dd9-7987-42a9-84cc-65dc912a6345"
   },
   "source": [
    "And now we can use our `make_predictions()` function to predict on `test_samples`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 142,
   "id": "79de2ac1-7d4b-4f81-ae8a-90099bca2a3d",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "79de2ac1-7d4b-4f81-ae8a-90099bca2a3d",
    "outputId": "918b07bc-4545-4401-84d5-8796ff5acf4c"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tensor([[3.9703e-07, 3.2616e-08, 2.5350e-09, 4.4762e-08, 6.7006e-09, 9.9961e-01,\n",
       "         5.7583e-07, 2.6825e-06, 2.5588e-05, 3.5969e-04],\n",
       "        [2.0484e-01, 5.9130e-01, 1.7537e-03, 1.2997e-01, 4.7323e-02, 3.4075e-05,\n",
       "         2.4563e-02, 1.5527e-04, 3.6486e-05, 1.8515e-05],\n",
       "        [2.5126e-05, 2.1583e-07, 6.0231e-06, 3.5479e-05, 1.1934e-06, 1.3174e-03,\n",
       "         3.3738e-05, 7.2422e-01, 2.4173e-03, 2.7194e-01],\n",
       "        [6.8008e-04, 3.5433e-04, 6.2171e-02, 2.3194e-04, 9.3581e-01, 1.0440e-06,\n",
       "         7.2480e-04, 3.0081e-06, 1.6470e-05, 7.0570e-06],\n",
       "        [3.1864e-02, 1.2982e-05, 3.9847e-05, 9.6254e-01, 1.1846e-04, 7.1308e-06,\n",
       "         5.3684e-03, 3.0355e-05, 1.4087e-05, 9.2336e-06],\n",
       "        [6.7622e-01, 2.7855e-05, 7.1031e-04, 1.6421e-03, 2.5809e-03, 3.5883e-07,\n",
       "         3.1844e-01, 2.4497e-06, 3.7496e-04, 2.7611e-07],\n",
       "        [1.9561e-04, 1.1711e-04, 4.4596e-02, 6.7087e-06, 9.3958e-01, 9.7547e-06,\n",
       "         1.5009e-02, 4.9496e-07, 4.2505e-04, 5.6911e-05],\n",
       "        [1.5457e-05, 1.9390e-05, 1.3786e-05, 1.5226e-04, 7.8964e-06, 3.3707e-03,\n",
       "         4.1528e-05, 9.9506e-01, 6.7065e-04, 6.4584e-04],\n",
       "        [7.9517e-06, 9.9918e-01, 1.1058e-08, 6.3306e-04, 1.7298e-04, 3.0249e-07,\n",
       "         1.8925e-06, 1.4483e-06, 4.1162e-07, 5.7998e-08]])"
      ]
     },
     "execution_count": 142,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Make predictions on test samples with model 2\n",
    "pred_probs= make_predictions(model=model_2, \n",
    "                             data=test_samples)\n",
    "\n",
    "# View first two prediction probabilities list\n",
    "pred_probs[:9]"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "22d3c080-4eb6-4b5d-a5c4-2319e78228af",
   "metadata": {
    "id": "22d3c080-4eb6-4b5d-a5c4-2319e78228af"
   },
   "source": [
    "Excellent!\n",
    "\n",
    "And now we can go from prediction probabilities to prediction labels by taking the `torch.argmax()` of the output of the `torch.softmax()` activation function."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 120,
   "id": "f9d97bcc-4310-4851-a1f8-6bcd757e9b26",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "f9d97bcc-4310-4851-a1f8-6bcd757e9b26",
    "outputId": "9d0f0bf9-a641-45e7-af77-6621fd1cfcc4"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tensor([5, 1, 9, 4, 3, 0, 4, 7, 1, 9, 7, 2, 0, 1, 0, 3])"
      ]
     },
     "execution_count": 120,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Turn the prediction probabilities into prediction labels by taking the argmax()\n",
    "pred_classes = pred_probs.argmax(dim=1)\n",
    "pred_classes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 121,
   "id": "1141af97-0990-4920-83d4-c13cca3f9abc",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "1141af97-0990-4920-83d4-c13cca3f9abc",
    "outputId": "c69cddd4-bbe9-495e-d477-6ea0a6c7d8de"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "([5, 1, 7, 4, 3, 0, 4, 7, 1, 9, 7, 2, 0, 1, 0, 4],\n",
       " tensor([5, 1, 9, 4, 3, 0, 4, 7, 1, 9, 7, 2, 0, 1, 0, 3]))"
      ]
     },
     "execution_count": 121,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Are our predictions in the same form as our test labels? \n",
    "test_labels, pred_classes"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4ea04387-c9ad-424f-8297-defd7b685683",
   "metadata": {
    "id": "4ea04387-c9ad-424f-8297-defd7b685683"
   },
   "source": [
    "Now our predicted classes are in the same format as our test labels, we can compare.\n",
    "\n",
    "Since we're dealing with image data, let's stay true to the data explorer's motto. \n",
    "\n",
    "\"Visualize, visualize, visualize!\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 122,
   "id": "87c8ac75",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['T-shirt/top',\n",
       " 'Trouser',\n",
       " 'Pullover',\n",
       " 'Dress',\n",
       " 'Coat',\n",
       " 'Sandal',\n",
       " 'Shirt',\n",
       " 'Sneaker',\n",
       " 'Bag',\n",
       " 'Ankle boot']"
      ]
     },
     "execution_count": 122,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "class_names"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 123,
   "id": "679cb5f7-bb66-42dd-a4d6-400b27b7c019",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 749
    },
    "id": "679cb5f7-bb66-42dd-a4d6-400b27b7c019",
    "outputId": "3aae0abe-9c19-4054-d8db-7e00403666aa"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1600x1600 with 16 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Plot predictions\n",
    "import numpy as np\n",
    "nrows = int(np.sqrt(N_SAMPLES))\n",
    "ncols = int(np.sqrt(N_SAMPLES))\n",
    "plt.figure(figsize=(16, 16))\n",
    "\n",
    "for i, sample in enumerate(test_samples):\n",
    "  # Create a subplot\n",
    "  plt.subplot(nrows, ncols, i+1)\n",
    "\n",
    "  # Plot the target image\n",
    "  plt.imshow(sample.squeeze(), cmap=\"gray\")\n",
    "\n",
    "  # Find the prediction label (in text form, e.g. \"Sandal\")\n",
    "  pred_label = class_names[pred_classes[i]]\n",
    "\n",
    "  # Get the truth label (in text form, e.g. \"T-shirt\")\n",
    "  truth_label = class_names[test_labels[i]] \n",
    "\n",
    "  # Create the title text of the plot\n",
    "  title_text = f\"Pred: {pred_label} | Truth: {truth_label}\"\n",
    "  \n",
    "  # Check for equality and change title colour accordingly\n",
    "  if pred_label == truth_label:\n",
    "      plt.title(title_text, fontsize=10, c=\"g\") # green text if correct\n",
    "  else:\n",
    "      plt.title(title_text, fontsize=10, c=\"r\") # red text if wrong\n",
    "  plt.axis(False);"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5ce6dc44-90a5-48c3-91a5-810fa084d98b",
   "metadata": {
    "id": "5ce6dc44-90a5-48c3-91a5-810fa084d98b"
   },
   "source": [
    "Well, well, well, doesn't that look good!\n",
    "\n",
    "Not bad for a couple dozen lines of PyTorch code!"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ab108078-6770-4cb9-ac62-a761ff159aba",
   "metadata": {
    "id": "ab108078-6770-4cb9-ac62-a761ff159aba"
   },
   "source": [
    "## 10. Making a confusion matrix for further prediction evaluation\n",
    "\n",
    "There are many [different evaluation metrics](https://www.learnpytorch.io/02_pytorch_classification/#9-more-classification-evaluation-metrics) we can use for classification problems. \n",
    "\n",
    "One of the most visual is a [confusion matrix](https://www.dataschool.io/simple-guide-to-confusion-matrix-terminology/).\n",
    "\n",
    "A confusion matrix shows you where your classification model got confused between predicitons and true labels.\n",
    "\n",
    "To make a confusion matrix, we'll go through three steps:\n",
    "1. Make predictions with our trained model, `model_2` (a confusion matrix compares predictions to true labels).\n",
    "2. Make a confusion matrix using [`torchmetrics.ConfusionMatrix`](https://torchmetrics.readthedocs.io/en/latest/references/modules.html?highlight=confusion#confusionmatrix). https://pytorch.org/ignite/generated/ignite.metrics.confusion_matrix.ConfusionMatrix.html\n",
    "3. Plot the confusion matrix using [`mlxtend.plotting.plot_confusion_matrix()`](http://rasbt.github.io/mlxtend/user_guide/plotting/plot_confusion_matrix/).\n",
    "\n",
    "Let's start by making predictions with our trained model."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 147,
   "id": "065b8090-c9c5-43df-b5c1-b45ba33af1be",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 49,
     "referenced_widgets": [
      "d3ab200da5f940d5b45396f83bd835e2",
      "f35a13b3e55342aeb24b188c1d81a9e5",
      "4a282c1974524bd3a7eba45fd3112129",
      "44d4196e99a4412f893ba8ac4672915d",
      "12d1a54d4107428eae2e64ff0a255c50",
      "4d6eb654b2794b0a95f31ac94b52a4ca",
      "fe5cff037f714657996f0541baee39f3",
      "0670e3e758e6486b9cf4e2797b4b619a",
      "3c590fc27b624584ba564e18bc42a2e4",
      "629ca5b704b84a958d4ee477907f64a1",
      "4d7c25dcdde8414382be4cf63a9cacf9"
     ]
    },
    "id": "065b8090-c9c5-43df-b5c1-b45ba33af1be",
    "outputId": "92a8bee2-71f5-4504-d534-cc63138c413d"
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Making predictions: 100%|██████████| 313/313 [00:01<00:00, 180.39it/s]\n"
     ]
    }
   ],
   "source": [
    "# Import tqdm for progress bar\n",
    "from tqdm.auto import tqdm\n",
    "\n",
    "# 1. Make predictions with trained model\n",
    "y_preds = []\n",
    "model_2.eval()\n",
    "with torch.inference_mode():\n",
    "  for X, y in tqdm(test_dataloader, desc=\"Making predictions\"):\n",
    "    # Send data and targets to target device\n",
    "    X, y = X.to(device), y.to(device)\n",
    "    # Do the forward pass\n",
    "    y_logit = model_2(X)\n",
    "    # Turn predictions from logits -> prediction probabilities -> predictions labels\n",
    "    y_pred = torch.softmax(y_logit, dim=1).argmax(dim=1) # note: perform softmax on the \"logits\" dimension, not \"batch\" dimension (in this case we have a batch size of 32, so can perform on dim=1)\n",
    "    # Put predictions on CPU for evaluation\n",
    "    y_preds.append(y_pred.cpu())\n",
    "# Concatenate list of predictions into a tensor\n",
    "y_pred_tensor = torch.cat(y_preds)\n",
    "#print(y_preds[:10].shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "362002d9-ec41-4c74-a210-b5d4f53410c4",
   "metadata": {
    "id": "362002d9-ec41-4c74-a210-b5d4f53410c4"
   },
   "source": [
    "Wonderful!\n",
    "\n",
    "Now we've got predictions, let's go through steps 2 & 3:\n",
    "2. Make a confusion matrix using [`torchmetrics.ConfusionMatrix`](https://torchmetrics.readthedocs.io/en/latest/references/modules.html?highlight=confusion#confusionmatrix).\n",
    "3. Plot the confusion matrix using [`mlxtend.plotting.plot_confusion_matrix()`](http://rasbt.github.io/mlxtend/user_guide/plotting/plot_confusion_matrix/).\n",
    "\n",
    "First we'll need to make sure we've got `torchmetrics` and `mlxtend` installed (these two libraries will help us make and visual a confusion matrix).\n",
    "\n",
    "> **Note:** If you're using Google Colab, the default version of `mlxtend` installed is 0.14.0 (as of March 2022), however, for the parameters of the `plot_confusion_matrix()` function we'd like use, we need 0.19.0 or higher. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 148,
   "id": "8c88c394",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "mlxtend version: 0.23.1\n"
     ]
    }
   ],
   "source": [
    "# See if torchmetrics exists, if not, install it\n",
    "import sys\n",
    "try:\n",
    "    import torchmetrics, mlxtend\n",
    "    print(f\"mlxtend version: {mlxtend.__version__}\")\n",
    "    assert int(mlxtend.__version__.split(\".\")[1]) >= 19, \"mlxtend version should be 0.19.0 or higher\"\n",
    "except:\n",
    "    # Use subprocess to run pip install in Python\n",
    "    import subprocess\n",
    "    subprocess.check_call([sys.executable, \"-m\", \"pip\", \"install\", \"--quiet\", \"torchmetrics\", \"-U\", \"mlxtend\"])\n",
    "    \n",
    "    import torchmetrics, mlxtend\n",
    "    print(f\"mlxtend version: {mlxtend.__version__}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 149,
   "id": "e6c0a05d-d3e0-4b86-9ef7-ee6ea5629b07",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "e6c0a05d-d3e0-4b86-9ef7-ee6ea5629b07",
    "outputId": "b37df16c-c292-4347-807c-91c97bf81f20"
   },
   "outputs": [],
   "source": [
    "# See if torchmetrics exists, if not, install it\n",
    "#try:\n",
    "#    import torchmetrics, mlxtend\n",
    "#    print(f\"mlxtend version: {mlxtend.__version__}\")\n",
    "#    assert int(mlxtend.__version__.split(\".\")[1]) >= 19, \"mlxtend verison should be 0.19.0 or higher\"\n",
    "#except:\n",
    "#    !pip install -q torchmetrics -U mlxtend # <- Note: If you're using Google Colab, this may require restarting the runtime\n",
    "#    import torchmetrics, mlxtend\n",
    "#    print(f\"mlxtend version: {mlxtend.__version__}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5245ede6-fd7f-40ad-a0b3-ae678544b84a",
   "metadata": {
    "id": "5245ede6-fd7f-40ad-a0b3-ae678544b84a"
   },
   "source": [
    "To plot the confusion matrix, we need to make sure we've got and [`mlxtend`](http://rasbt.github.io/mlxtend/) version of 0.19.0 or higher."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 150,
   "id": "21383f88-a2dd-4678-94c6-479c592da0ab",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "21383f88-a2dd-4678-94c6-479c592da0ab",
    "outputId": "faffbe4c-9c86-4a20-cbd6-c7e8e48e81a5"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.23.1\n"
     ]
    }
   ],
   "source": [
    "# Import mlxtend upgraded version\n",
    "import mlxtend \n",
    "print(mlxtend.__version__)\n",
    "assert int(mlxtend.__version__.split(\".\")[1]) >= 19 # should be version 0.19.0 or higher"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c91b9346-e25f-48ab-967e-425649331dc6",
   "metadata": {
    "id": "c91b9346-e25f-48ab-967e-425649331dc6"
   },
   "source": [
    "`torchmetrics` and `mlxtend` installed, let's make a confusion matrix!\n",
    "\n",
    "First we'll create a `torchmetrics.ConfusionMatrix` instance telling it how many classes we're dealing with by setting `num_classes=len(class_names)`.\n",
    "\n",
    "Then we'll create a confusion matrix (in tensor format) by passing our instance our model's predictions (`preds=y_pred_tensor`) and targets (`target=test_data.targets`).\n",
    "\n",
    "Finally we can plot our confision matrix using the `plot_confusion_matrix()` function from `mlxtend.plotting`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 151,
   "id": "7aed6d76-ad1c-429e-b8e0-c80572e3ebf4",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 667
    },
    "id": "7aed6d76-ad1c-429e-b8e0-c80572e3ebf4",
    "outputId": "ae34ae74-2038-4037-f01d-77a807e4de9b"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1000x700 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from torchmetrics import ConfusionMatrix\n",
    "from mlxtend.plotting import plot_confusion_matrix\n",
    "\n",
    "# 2. Setup confusion matrix instance and compare predictions to targets\n",
    "confmat = ConfusionMatrix(num_classes=len(class_names), task='multiclass')\n",
    "confmat_tensor = confmat(preds=y_pred_tensor,\n",
    "                         target=test_data.targets)\n",
    "\n",
    "# 3. Plot the confusion matrix\n",
    "fig, ax = plot_confusion_matrix(\n",
    "    conf_mat=confmat_tensor.numpy(), # matplotlib likes working with NumPy \n",
    "    class_names=class_names, # turn the row and column labels into class names\n",
    "    figsize=(10, 7)\n",
    ");"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "381c1c93-df30-451c-b65e-5d4c1680dc30",
   "metadata": {
    "id": "381c1c93-df30-451c-b65e-5d4c1680dc30"
   },
   "source": [
    "Woah! Doesn't that look good?\n",
    "\n",
    "We can see our model does fairly well since most of the dark squares are down the diagonal from top left to bottom right (and ideal model will have only values in these squares and 0 everywhere else).\n",
    "\n",
    "The model gets most \"confused\" on classes that are similar, for example predicting \"Pullover\" for images that are actually labelled \"Shirt\".\n",
    "\n",
    "And the same for predicting \"Shirt\" for classes that are actually labelled \"T-shirt/top\".\n",
    "\n",
    "This kind of information is often more helpful than a single accuracy metric because it tells use *where* a model is getting things wrong.\n",
    "\n",
    "It also hints at *why* the model may be getting certain things wrong.\n",
    "\n",
    "It's understandable the model sometimes predicts \"Shirt\" for images labelled \"T-shirt/top\".\n",
    "\n",
    "We can use this kind of information to further inspect our models and data to see how it could be improved.\n",
    "\n",
    "> **Exercise:** Use the trained `model_2` to make predictions on the test FashionMNIST dataset. Then plot some predictions where the model was wrong alongside what the label of the image should've been. After visualing these predictions do you think it's more of a modelling error or a data error? As in, could the model do better or are the labels of the data too close to each other (e.g. a \"Shirt\" label is too close to \"T-shirt/top\")?"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "25818e83-89de-496d-8b56-af4fc9f2acc5",
   "metadata": {
    "id": "25818e83-89de-496d-8b56-af4fc9f2acc5"
   },
   "source": [
    "## 11. Save and load best performing model\n",
    "\n",
    "Let's finish this section off by saving and loading in our best performing model.\n",
    "\n",
    "Recall from [notebook 01](https://www.learnpytorch.io/01_pytorch_workflow/#5-saving-and-loading-a-pytorch-model) we can save and load a PyTorch model using a combination of:\n",
    "* `torch.save` - a function to save a whole PyTorch model or a model's `state_dict()`. \n",
    "* `torch.load` - a function to load in a saved PyTorch object.\n",
    "* `torch.nn.Module.load_state_dict()` - a function to load a saved `state_dict()` into an existing model instance.\n",
    "\n",
    "You can see more of these three in the [PyTorch saving and loading models documentation](https://pytorch.org/tutorials/beginner/saving_loading_models.html).\n",
    "\n",
    "For now, let's save our `model_2`'s `state_dict()` then load it back in and evaluate it to make sure the save and load went correctly. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 152,
   "id": "d058e8fa-560f-4350-a154-49593ff403c9",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "d058e8fa-560f-4350-a154-49593ff403c9",
    "outputId": "0156a518-dae2-4b25-999a-c0a77ef7ef7c"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Saving model to: models\\03_pytorch_computer_vision_model_2.pth\n"
     ]
    }
   ],
   "source": [
    "from pathlib import Path\n",
    "\n",
    "# Create models directory (if it doesn't already exist), see: https://docs.python.org/3/library/pathlib.html#pathlib.Path.mkdir\n",
    "MODEL_PATH = Path(\"models\")\n",
    "MODEL_PATH.mkdir(parents=True, # create parent directories if needed\n",
    "                 exist_ok=True # if models directory already exists, don't error\n",
    ")\n",
    "\n",
    "# Create model save path\n",
    "MODEL_NAME = \"03_pytorch_computer_vision_model_2.pth\"\n",
    "MODEL_SAVE_PATH = MODEL_PATH / MODEL_NAME\n",
    "\n",
    "# Save the model state dict\n",
    "print(f\"Saving model to: {MODEL_SAVE_PATH}\")\n",
    "torch.save(obj=model_2.state_dict(), # only saving the state_dict() only saves the learned parameters\n",
    "           f=MODEL_SAVE_PATH)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a1542284-8132-42ba-b00d-57e9b9037e4e",
   "metadata": {
    "id": "a1542284-8132-42ba-b00d-57e9b9037e4e"
   },
   "source": [
    "Now we've got a saved model `state_dict()` we can load it back in using a combination of `load_state_dict()` and `torch.load()`.\n",
    "\n",
    "Since we're using `load_state_dict()`, we'll need to create a new instance of `FashionMNISTModelV2()` with the same input parameters as our saved model `state_dict()`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 153,
   "id": "634a8f7a-3013-4b45-b365-49b286d3c478",
   "metadata": {
    "id": "634a8f7a-3013-4b45-b365-49b286d3c478"
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\ssre_\\AppData\\Local\\Temp\\ipykernel_30816\\3689379356.py:8: FutureWarning: You are using `torch.load` with `weights_only=False` (the current default value), which uses the default pickle module implicitly. It is possible to construct malicious pickle data which will execute arbitrary code during unpickling (See https://github.com/pytorch/pytorch/blob/main/SECURITY.md#untrusted-models for more details). In a future release, the default value for `weights_only` will be flipped to `True`. This limits the functions that could be executed during unpickling. Arbitrary objects will no longer be allowed to be loaded via this mode unless they are explicitly allowlisted by the user via `torch.serialization.add_safe_globals`. We recommend you start setting `weights_only=True` for any use case where you don't have full control of the loaded file. Please open an issue on GitHub for any issues related to this experimental feature.\n",
      "  loaded_model_2.load_state_dict(torch.load(f=MODEL_SAVE_PATH))\n"
     ]
    }
   ],
   "source": [
    "# Create a new instance of FashionMNISTModelV2 (the same class as our saved state_dict())\n",
    "# Note: loading model will error if the shapes here aren't the same as the saved version\n",
    "loaded_model_2 = FashionMNISTModelV2(input_shape=1, \n",
    "                                    hidden_units=10, # try changing this to 128 and seeing what happens \n",
    "                                    output_shape=10) \n",
    "\n",
    "# Load in the saved state_dict()\n",
    "loaded_model_2.load_state_dict(torch.load(f=MODEL_SAVE_PATH))\n",
    "\n",
    "# Send model to GPU\n",
    "loaded_model_2 = loaded_model_2.to(device)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "feeaebf4-6040-4fa5-852d-5eb8d2bbb94c",
   "metadata": {
    "id": "feeaebf4-6040-4fa5-852d-5eb8d2bbb94c"
   },
   "source": [
    "And now we've got a loaded model we can evaluate it with `eval_model()` to make sure its parameters work similarly to `model_2` prior to saving. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 154,
   "id": "3e3bcd06-d99b-47bc-8828-9e3903285599",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "3e3bcd06-d99b-47bc-8828-9e3903285599",
    "outputId": "c0ee1d5f-9573-4e1a-8430-ee09fb4d72cd"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'model_name': 'FashionMNISTModelV2',\n",
       " 'model_loss': 0.31258341670036316,\n",
       " 'model_acc': 88.87779552715655}"
      ]
     },
     "execution_count": 154,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Evaluate loaded model\n",
    "torch.manual_seed(42)\n",
    "\n",
    "loaded_model_2_results = eval_model(\n",
    "    model=loaded_model_2,\n",
    "    data_loader=test_dataloader,\n",
    "    loss_fn=loss_fn, \n",
    "    accuracy_fn=accuracy_fn\n",
    ")\n",
    "\n",
    "loaded_model_2_results"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c2b37855-c0da-4834-a2d4-a0faa8410b65",
   "metadata": {
    "id": "c2b37855-c0da-4834-a2d4-a0faa8410b65"
   },
   "source": [
    "Do these results look the same as `model_2_results`?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 155,
   "id": "68544254-c99a-47ec-a32f-9816c21a993e",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "68544254-c99a-47ec-a32f-9816c21a993e",
    "outputId": "74b8d4ca-d35a-4f70-e8b9-ed54f034358e"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'model_name': 'FashionMNISTModelV2',\n",
       " 'model_loss': 0.31258341670036316,\n",
       " 'model_acc': 88.87779552715655}"
      ]
     },
     "execution_count": 155,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model_2_results"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0ee07f93-4344-4c7a-8b1d-92a56034e7b2",
   "metadata": {
    "id": "0ee07f93-4344-4c7a-8b1d-92a56034e7b2"
   },
   "source": [
    "We can find out if two tensors are close to each other using `torch.isclose()` and passing in a tolerance level of closeness via the parameters `atol` (absolute tolerance) and `rtol` (relative tolerance).\n",
    "\n",
    "If our model's results are close, the output of `torch.isclose()` should be true."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 156,
   "id": "48dcf0ba-7e00-4406-8aaa-41918856361a",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "48dcf0ba-7e00-4406-8aaa-41918856361a",
    "outputId": "47324300-0d00-46de-d130-1283ad044ef8"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tensor(True)"
      ]
     },
     "execution_count": 156,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Check to see if results are close to each other (if they are very far away, there may be an error)\n",
    "torch.isclose(torch.tensor(model_2_results[\"model_loss\"]), \n",
    "              torch.tensor(loaded_model_2_results[\"model_loss\"]),\n",
    "              atol=1e-08, # absolute tolerance\n",
    "              rtol=0.0001) # relative tolerance"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c3969b7d-9955-4b6f-abf8-fe8eedf233a9",
   "metadata": {
    "id": "c3969b7d-9955-4b6f-abf8-fe8eedf233a9"
   },
   "source": [
    "## Exercises\n",
    "\n",
    "All of the exercises are focused on practicing the code in the sections above.\n",
    "\n",
    "You should be able to complete them by referencing each section or by following the resource(s) linked.\n",
    "\n",
    "All exercises should be completed using [device-agnostic code](https://pytorch.org/docs/stable/notes/cuda.html#device-agnostic-code).\n",
    "\n",
    "**Resources:**\n",
    "* [Exercise template notebook for 03](https://github.com/mrdbourke/pytorch-deep-learning/blob/main/extras/exercises/03_pytorch_computer_vision_exercises.ipynb)\n",
    "* [Example solutions notebook for 03](https://github.com/mrdbourke/pytorch-deep-learning/blob/main/extras/solutions/03_pytorch_computer_vision_exercise_solutions.ipynb) (try the exercises *before* looking at this)\n",
    "\n",
    "1. What are 3 areas in industry where computer vision is currently being used?\n",
    "2. Search \"what is overfitting in machine learning\" and write down a sentence about what you find. \n",
    "3. Search \"ways to prevent overfitting in machine learning\", write down 3 of the things you find and a sentence about each. **Note:** there are lots of these, so don't worry too much about all of them, just pick 3 and start with those.\n",
    "4. Spend 20-minutes reading and clicking through the [CNN Explainer website](https://poloclub.github.io/cnn-explainer/).\n",
    "    * Upload your own example image using the \"upload\" button and see what happens in each layer of a CNN as your image passes through it.\n",
    "5. Load the [`torchvision.datasets.MNIST()`](https://pytorch.org/vision/stable/generated/torchvision.datasets.MNIST.html#torchvision.datasets.MNIST) train and test datasets.\n",
    "6. Visualize at least 5 different samples of the MNIST training dataset.\n",
    "7. Turn the MNIST train and test datasets into dataloaders using `torch.utils.data.DataLoader`, set the `batch_size=32`.\n",
    "8. Recreate `model_2` used in this notebook (the same model from the [CNN Explainer website](https://poloclub.github.io/cnn-explainer/), also known as TinyVGG) capable of fitting on the MNIST dataset.\n",
    "9. Train the model you built in exercise 8. on CPU and GPU and see how long it takes on each.\n",
    "10. Make predictions using your trained model and visualize at least 5 of them comparing the prediciton to the target label.\n",
    "11. Plot a confusion matrix comparing your model's predictions to the truth labels.\n",
    "12. Create a random tensor of shape `[1, 3, 64, 64]` and pass it through a `nn.Conv2d()` layer with various hyperparameter settings (these can be any settings you choose), what do you notice if the `kernel_size` parameter goes up and down?\n",
    "13. Use a model similar to the trained `model_2` from this notebook to make predictions on the test [`torchvision.datasets.FashionMNIST`](https://pytorch.org/vision/main/generated/torchvision.datasets.FashionMNIST.html) dataset. \n",
    "    * Then plot some predictions where the model was wrong alongside what the label of the image should've been. \n",
    "    * After visualing these predictions do you think it's more of a modelling error or a data error? \n",
    "    * As in, could the model do better or are the labels of the data too close to each other (e.g. a \"Shirt\" label is too close to \"T-shirt/top\")?\n",
    "\n",
    "## Extra-curriculum\n",
    "* **Watch:** [MIT's Introduction to Deep Computer Vision](https://www.youtube.com/watch?v=iaSUYvmCekI&list=PLtBw6njQRU-rwp5__7C0oIVt26ZgjG9NI&index=3) lecture. This will give you a great intuition behind convolutional neural networks.\n",
    "* Spend 10-minutes clicking thorugh the different options of the [PyTorch vision library](https://pytorch.org/vision/stable/index.html), what different modules are available?\n",
    "* Lookup \"most common convolutional neural networks\", what architectures do you find? Are any of them contained within the [`torchvision.models`](https://pytorch.org/vision/stable/models.html) library? What do you think you could do with these?\n",
    "* For a large number of pretrained PyTorch computer vision models as well as many different extensions to PyTorch's computer vision functionalities check out the [PyTorch Image Models library `timm`](https://github.com/rwightman/pytorch-image-models/) (Torch Image Models) by Ross Wightman."
   ]
  }
 ],
 "metadata": {
  "accelerator": "GPU",
  "colab": {
   "gpuType": "A100",
   "machine_shape": "hm",
   "provenance": []
  },
  "gpuClass": "standard",
  "kernelspec": {
   "display_name": ".venv_pytorch_gpu",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.10.14"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}