examples/pytorch-ignite.ipynb

{
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "collapsed": false
      },
      "source": [
        "# livelossplot example: PyTorch Ignite\n",
        "\n",
        "Last update: `livelossplot 0.5.0`. For code and documentation, see [livelossplot GitHub repository](https://github.com/stared/livelossplot).\n",
        "\n",
        "<a href=\"https://colab.research.google.com/github/stared/livelossplot/blob/main/examples/pytorch-ignite.ipynb\" target=\"_parent\">\n",
        "    <img src=\"https://colab.research.google.com/assets/colab-badge.svg\"/>\n",
        "</a>\n",
        "\n",
        "## Note\n",
        "\n",
        "The syntax changed, and from `0.5+` on it is):\n",
        "\n",
        "```{python}\n",
        "from livelossplot import PlotLossesIgnite\n",
        "```"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "collapsed": false,
        "pycharm": {
          "name": "#%%\n"
        }
      },
      "outputs": [],
      "source": [
        "!pip install pytorch-ignite livelossplot --quiet"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "pycharm": {
          "is_executing": true
        }
      },
      "outputs": [],
      "source": [
        "%matplotlib inline\n",
        "\n",
        "from torch import nn\n",
        "import torch.optim\n",
        "import torch.nn.functional as F\n",
        "from torch.utils.data import DataLoader\n",
        "from torchvision.transforms import Compose, ToTensor, Normalize\n",
        "from torchvision.datasets import MNIST\n",
        "\n",
        "from ignite.engine import Events, create_supervised_trainer, create_supervised_evaluator\n",
        "from ignite.metrics import Accuracy, Loss\n",
        "from ignite.contrib.handlers import ProgressBar\n",
        "\n",
        "from livelossplot import PlotLossesIgnite "
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "pycharm": {
          "is_executing": true
        }
      },
      "outputs": [],
      "source": [
        "TRAIN_BATCH_SIZE = 100\n",
        "TEST_BATCH_SIZE = 100"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "pycharm": {
          "is_executing": true
        }
      },
      "outputs": [],
      "source": [
        "class Net(nn.Module):\n",
        "    def __init__(self):\n",
        "        super(Net, self).__init__()\n",
        "        self.conv1 = nn.Conv2d(1, 10, kernel_size=5)\n",
        "        self.conv2 = nn.Conv2d(10, 20, kernel_size=5)\n",
        "        self.conv2_drop = nn.Dropout2d()\n",
        "        self.fc1 = nn.Linear(320, 50)\n",
        "        self.fc2 = nn.Linear(50, 10)\n",
        "\n",
        "    def forward(self, x):\n",
        "        x = F.relu(F.max_pool2d(self.conv1(x), 2))\n",
        "        x = F.relu(F.max_pool2d(self.conv2_drop(self.conv2(x)), 2))\n",
        "        x = x.view(-1, 320)\n",
        "        x = F.relu(self.fc1(x))\n",
        "        x = F.dropout(x, training=self.training)\n",
        "        x = self.fc2(x)\n",
        "        return F.log_softmax(x, dim=-1)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "pycharm": {
          "is_executing": true
        }
      },
      "outputs": [],
      "source": [
        "def get_data_loaders(train_batch_size, val_batch_size):\n",
        "    data_transform = Compose([ToTensor(), Normalize((0.1307,), (0.3081,))])\n",
        "\n",
        "    train_loader = DataLoader(\n",
        "        MNIST(download=True, root=\".\", transform=data_transform, train=True), batch_size=train_batch_size, shuffle=True\n",
        "    )\n",
        "\n",
        "    val_loader = DataLoader(\n",
        "        MNIST(download=False, root=\".\", transform=data_transform, train=False), batch_size=val_batch_size, shuffle=False\n",
        "    )\n",
        "    return train_loader, val_loader"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "pycharm": {
          "is_executing": true
        }
      },
      "outputs": [],
      "source": [
        "model = Net()\n",
        "train_loader, val_loader = get_data_loaders(TRAIN_BATCH_SIZE, TEST_BATCH_SIZE)\n",
        "optimizer = torch.optim.SGD(model.parameters(), lr=0.01, momentum=0.8)\n",
        "loss = torch.nn.NLLLoss()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "pycharm": {
          "is_executing": true
        }
      },
      "outputs": [],
      "source": [
        "trainer = create_supervised_trainer(model, optimizer, loss)\n",
        "metrics = metrics={\n",
        "    'val_acc': Accuracy(),\n",
        "    'val_loss': Loss(loss)\n",
        "}\n",
        "p_bar = ProgressBar()\n",
        "p_bar.attach(trainer)\n",
        "evaluator = create_supervised_evaluator(model, metrics=metrics)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "pycharm": {
          "is_executing": true
        }
      },
      "outputs": [],
      "source": [
        "@trainer.on(Events.EPOCH_COMPLETED)\n",
        "def validate(trainer):\n",
        "    evaluator.run(val_loader)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "pycharm": {
          "is_executing": true
        }
      },
      "outputs": [],
      "source": [
        "callback = PlotLossesIgnite()\n",
        "callback.attach(evaluator)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "pycharm": {
          "is_executing": true,
          "name": "#%%\n"
        }
      },
      "outputs": [],
      "source": [
        "trainer.run(train_loader, max_epochs=100)"
      ]
    }
  ],
  "metadata": {
    "kernelspec": {
      "display_name": "Python 3",
      "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": "3.5.6"
    },
    "pycharm": {
      "stem_cell": {
        "cell_type": "raw",
        "source": [],
        "metadata": {
          "collapsed": false
        }
      }
    }
  },
  "nbformat": 4,
  "nbformat_minor": 1
}