Qkeras_sandbox3.ipynb

{"cells":[{"cell_type":"markdown","metadata":{"id":"6IAvKMEMlBN0"},"source":["# Connect to Drive"]},{"cell_type":"code","execution_count":null,"metadata":{"id":"wtczU_fMrGcq"},"outputs":[],"source":["import os\n","import sys"]},{"cell_type":"code","execution_count":null,"metadata":{"id":"xa3mNYpYkmeV"},"outputs":[],"source":["from google.colab import drive\n","MOUNTPOINT = '/content/gdrive'\n","DRIVE_PATH = os.path.join(MOUNTPOINT, 'My Drive')\n","drive.mount(MOUNTPOINT)\n","sys.path.append(os.path.join(DRIVE_PATH, 'matt_schoenbauer_qkeras_project'))\n","sys.path.append(os.path.join(DRIVE_PATH, 'matt_schoenbauer_qkeras_project', 'qkeras'))"]},{"cell_type":"code","execution_count":null,"metadata":{"id":"FLsEP5x0ny2-"},"outputs":[],"source":["%load_ext autoreload"]},{"cell_type":"code","execution_count":null,"metadata":{"id":"UHqxsyy2n1BB"},"outputs":[],"source":["%autoreload 1"]},{"cell_type":"code","execution_count":null,"metadata":{"id":"SAiRueNO2RgA"},"outputs":[],"source":["!pip install -r /content/gdrive/My\\ Drive/matt_schoenbauer_qkeras_project/qkeras/requirements.txt > /dev/null"]},{"cell_type":"code","execution_count":6,"metadata":{"id":"mQ4lYuifmem2","colab":{"base_uri":"https://localhost:8080/"},"executionInfo":{"status":"ok","timestamp":1679927836782,"user_tz":240,"elapsed":29167,"user":{"displayName":"Matt Schoenbauer","userId":"02579049095237637221"}},"outputId":"4705e4fe-7bef-49f9-d212-f7df091711be"},"outputs":[{"output_type":"stream","name":"stdout","text":["File timing_utils.py Updated. Last modified 4 days, 1:14:48 seconds ago\n","File correctness_utils.py Updated. Last modified 22:36:53 seconds ago\n","File alt_implementations.py Updated. Last modified 12:25:52 seconds ago\n","File alt_implementations2.py Updated. Last modified 4 days, 1:18:16 seconds ago\n"]}],"source":["from time import time\n","from importlib import reload\n","\n","import numpy as np\n","import pandas as pd\n","import tensorflow as tf\n","from matplotlib import pyplot as plt\n","\n","%aimport timing_utils\n","%aimport correctness_utils \n","%aimport alt_implementations\n","%aimport alt_implementations2\n","\n","%aimport qkeras"]},{"cell_type":"markdown","metadata":{"id":"Uw9YHplrHsih"},"source":["# Correctness tests"]},{"cell_type":"code","execution_count":7,"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"executionInfo":{"elapsed":5,"status":"ok","timestamp":1679927843570,"user":{"displayName":"Matt Schoenbauer","userId":"02579049095237637221"},"user_tz":240},"id":"LkSbsW7SjZWd","outputId":"9b293bbb-b6a6-4eeb-d3a1-e926b2a8924d"},"outputs":[{"output_type":"execute_result","data":{"text/plain":["1"]},"metadata":{},"execution_count":7}],"source":["1"]},{"cell_type":"code","execution_count":null,"metadata":{"colab":{"background_save":true,"base_uri":"https://localhost:8080/"},"id":"Q2iUIkB32WQx","outputId":"ba84378a-4ec1-40f9-b50f-e2c535165111"},"outputs":[{"name":"stderr","output_type":"stream","text":["Accuracy tests (linear):   0%|          | 0/24 [00:00<?, ?it/s]WARNING:tensorflow:5 out of the last 5 calls to <function quantized_bits2.__call__ at 0x7f4dfbe17310> triggered tf.function retracing. Tracing is expensive and the excessive number of tracings could be due to (1) creating @tf.function repeatedly in a loop, (2) passing tensors with different shapes, (3) passing Python objects instead of tensors. For (1), please define your @tf.function outside of the loop. For (2), @tf.function has reduce_retracing=True option that can avoid unnecessary retracing. For (3), please refer to https://www.tensorflow.org/guide/function#controlling_retracing and https://www.tensorflow.org/api_docs/python/tf/function for  more details.\n","WARNING:tensorflow:6 out of the last 6 calls to <function quantized_bits2.__call__ at 0x7f4dfbe17310> triggered tf.function retracing. Tracing is expensive and the excessive number of tracings could be due to (1) creating @tf.function repeatedly in a loop, (2) passing tensors with different shapes, (3) passing Python objects instead of tensors. For (1), please define your @tf.function outside of the loop. For (2), @tf.function has reduce_retracing=True option that can avoid unnecessary retracing. For (3), please refer to https://www.tensorflow.org/guide/function#controlling_retracing and https://www.tensorflow.org/api_docs/python/tf/function for  more details.\n","Accuracy tests (linear):  83%|████████▎ | 20/24 [11:39<02:19, 34.96s/it]\n"]},{"ename":"AssertionError","evalue":"ignored","output_type":"error","traceback":["\u001b[0;31m---------------------------------------------------------------------------\u001b[0m","\u001b[0;31mAssertionError\u001b[0m                            Traceback (most recent call last)","\u001b[0;32m<ipython-input-172-d43184271574>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m      1\u001b[0m \u001b[0mmax_test_count\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;36m50\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 2\u001b[0;31m \u001b[0mcorrectness_utils\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mquantized_bits_linear_accuracy_tests\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0malt_implementations\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mquantized_bits2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      3\u001b[0m \u001b[0mcorrectness_utils\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mquantized_bits_grid_accuracy_tests\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0malt_implementations\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mquantized_bits2\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mmax_test_count\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mmax_test_count\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n","\u001b[0;32m/content/gdrive/My Drive/matt_schoenbauer_qkeras_project/correctness_utils.py\u001b[0m in \u001b[0;36mquantized_bits_linear_accuracy_tests\u001b[0;34m(alt_quantized_bits)\u001b[0m\n\u001b[1;32m     80\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     81\u001b[0m     \u001b[0;32mfor\u001b[0m \u001b[0mkwargs\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mtqdm\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mkwargs_list\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdesc\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m\"Accuracy tests (linear)\"\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 82\u001b[0;31m         \u001b[0m_check_quantized_bits_correctness\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0malt_quantized_bits\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     83\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     84\u001b[0m     \u001b[0mprint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\"\\nLinear accuracy tests passed!\\n\"\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n","\u001b[0;32m/content/gdrive/My Drive/matt_schoenbauer_qkeras_project/correctness_utils.py\u001b[0m in \u001b[0;36m_check_quantized_bits_correctness\u001b[0;34m(alt_quantized_bits, kwargs)\u001b[0m\n\u001b[1;32m    108\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    109\u001b[0m     \u001b[0;32mfor\u001b[0m \u001b[0mx\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mTEST_X_VALUES\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 110\u001b[0;31m         \u001b[0m_check_correctness\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0malt\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mbaseline\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mx\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    111\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    112\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n","\u001b[0;32m/content/gdrive/My Drive/matt_schoenbauer_qkeras_project/correctness_utils.py\u001b[0m in \u001b[0;36m_check_correctness\u001b[0;34m(alt_func, baseline_func, x, kwargs)\u001b[0m\n\u001b[1;32m    128\u001b[0m         \u001b[0;31m# import pdb; pdb.set_trace()\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    129\u001b[0m     \u001b[0;32mif\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0mnp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mallclose\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mbaseline_scale\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0malt_scale\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mand\u001b[0m \u001b[0mK\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmax\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m>\u001b[0m \u001b[0;36m0\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 130\u001b[0;31m         \u001b[0;32massert\u001b[0m \u001b[0;32mFalse\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0merr_msg\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    131\u001b[0m         \u001b[0;31m# import pdb; pdb.set_trace()\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    132\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n","\u001b[0;31mAssertionError\u001b[0m: Failed for {'alpha': 'auto', 'symmetric': True, 'keep_negative': True} with x = 0.04081632653061229. \nbaseline_res = 0.040816326, alt_res = 0.040816326. \nbaseline_scale = array(0.04113771, dtype=float32), alt_scale = array(1., dtype=float32)"]}],"source":["max_test_count = 50\n","correctness_utils.quantized_bits_linear_accuracy_tests(alt_implementations.quantized_bits2)\n","correctness_utils.quantized_bits_grid_accuracy_tests(alt_implementations.quantized_bits2, max_test_count=max_test_count)\n"]},{"cell_type":"code","execution_count":123,"metadata":{"id":"v7QH0rR0YGiP","executionInfo":{"status":"ok","timestamp":1679935271312,"user_tz":240,"elapsed":6,"user":{"displayName":"Matt Schoenbauer","userId":"02579049095237637221"}}},"outputs":[],"source":["\n","# Debugging code for correctness tests\n","# This works when the noise factor is zero, but not otherwise\n","\n","# kwargs = {'alpha': 'auto', 'bits': 4, 'integer': 1, 'symmetric': True, \n","#           'keep_negative': True, 'qnoise_factor': 0.0, 'use_stochastic_rounding': False, 'use_ste': True}\n","kwargs = {'alpha': 'auto', 'symmetric': True, 'keep_negative': True}\n","kwargs = {}\n","baseline = correctness_utils.quantized_bits(**kwargs)\n","alt = alt_implementations.quantized_bits2(**kwargs)\n","stupid = quantized_bits_stupid(**{'alpha': 'auto'})\n","x = tf.random.normal((1,))\n","\n"]},{"cell_type":"code","source":["x"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"vQ188DtC8htB","executionInfo":{"status":"ok","timestamp":1679935285608,"user_tz":240,"elapsed":6,"user":{"displayName":"Matt Schoenbauer","userId":"02579049095237637221"}},"outputId":"7dc97019-0843-46f4-bec4-b7f006afac4a"},"execution_count":125,"outputs":[{"output_type":"execute_result","data":{"text/plain":["<tf.Tensor: shape=(1,), dtype=float32, numpy=array([0.6332401], dtype=float32)>"]},"metadata":{},"execution_count":125}]},{"cell_type":"code","source":["baseline_res = baseline(x).numpy()\n","alt_res = alt(x).numpy()\n","# stupid_res = stupid(x).numpy()\n","\n","print(alt_res, baseline_res)\n","baseline.scale, alt.scale\n","\n","# print(stupid_res, baseline_res)\n","# baseline.scale, stupid.scale"],"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":669},"id":"fWWbTZZKumw_","executionInfo":{"status":"error","timestamp":1679935273366,"user_tz":240,"elapsed":9,"user":{"displayName":"Matt Schoenbauer","userId":"02579049095237637221"}},"outputId":"6f7d2883-73bf-40ed-d1f3-cd7499f94c1d"},"execution_count":124,"outputs":[{"output_type":"error","ename":"ValueError","evalue":"ignored","traceback":["\u001b[0;31m---------------------------------------------------------------------------\u001b[0m","\u001b[0;31mValueError\u001b[0m                                Traceback (most recent call last)","\u001b[0;32m<ipython-input-124-4c54bb4da2c7>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m      1\u001b[0m \u001b[0mbaseline_res\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mbaseline\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mnumpy\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 2\u001b[0;31m \u001b[0malt_res\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0malt\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mnumpy\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      3\u001b[0m \u001b[0;31m# stupid_res = stupid(x).numpy()\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      4\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      5\u001b[0m \u001b[0mprint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0malt_res\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mbaseline_res\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n","\u001b[0;32m/usr/local/lib/python3.9/dist-packages/tensorflow/python/util/traceback_utils.py\u001b[0m in \u001b[0;36merror_handler\u001b[0;34m(*args, **kwargs)\u001b[0m\n\u001b[1;32m    151\u001b[0m     \u001b[0;32mexcept\u001b[0m \u001b[0mException\u001b[0m \u001b[0;32mas\u001b[0m \u001b[0me\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    152\u001b[0m       \u001b[0mfiltered_tb\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0m_process_traceback_frames\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0me\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m__traceback__\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 153\u001b[0;31m       \u001b[0;32mraise\u001b[0m \u001b[0me\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mwith_traceback\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfiltered_tb\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mfrom\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    154\u001b[0m     \u001b[0;32mfinally\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    155\u001b[0m       \u001b[0;32mdel\u001b[0m \u001b[0mfiltered_tb\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n","\u001b[0;32m/content/gdrive/My Drive/matt_schoenbauer_qkeras_project/alt_implementations.py\u001b[0m in \u001b[0;36mtf____call__\u001b[0;34m(self, x)\u001b[0m\n\u001b[1;32m     12\u001b[0m                 \u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mconverted_call\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_build\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfscope\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     13\u001b[0m                 \u001b[0mx\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mconverted_call\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mK\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcast_to_floatx\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfscope\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 14\u001b[0;31m                 \u001b[0mxq\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mconverted_call\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtf\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcond\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0muse_sign_function\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mautograph_artifact\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;32mlambda\u001b[0m \u001b[0;34m:\u001b[0m \u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mconverted_call\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_sign_function\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfscope\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mautograph_artifact\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;32mlambda\u001b[0m \u001b[0;34m:\u001b[0m \u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mconverted_call\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_multi_bit_computation\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfscope\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfscope\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     15\u001b[0m                 \u001b[0;32mtry\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     16\u001b[0m                     \u001b[0mdo_return\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;32mTrue\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n","\u001b[0;32m/content/gdrive/My Drive/matt_schoenbauer_qkeras_project/alt_implementations.py\u001b[0m in \u001b[0;36m<lambda>\u001b[0;34m()\u001b[0m\n\u001b[1;32m     12\u001b[0m                 \u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mconverted_call\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_build\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfscope\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     13\u001b[0m                 \u001b[0mx\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mconverted_call\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mK\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcast_to_floatx\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfscope\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 14\u001b[0;31m                 \u001b[0mxq\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mconverted_call\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtf\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcond\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0muse_sign_function\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mautograph_artifact\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;32mlambda\u001b[0m \u001b[0;34m:\u001b[0m \u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mconverted_call\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_sign_function\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfscope\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mautograph_artifact\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;32mlambda\u001b[0m \u001b[0;34m:\u001b[0m \u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mconverted_call\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_multi_bit_computation\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfscope\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfscope\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     15\u001b[0m                 \u001b[0;32mtry\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     16\u001b[0m                     \u001b[0mdo_return\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;32mTrue\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n","\u001b[0;32m/content/gdrive/My Drive/matt_schoenbauer_qkeras_project/alt_implementations.py\u001b[0m in \u001b[0;36mtf___multi_bit_computation\u001b[0;34m(self, x)\u001b[0m\n\u001b[1;32m     36\u001b[0m                             \u001b[0;32mraise\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     37\u001b[0m                         \u001b[0;32mreturn\u001b[0m \u001b[0mfscope_2\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mret\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mretval__2\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdo_return_2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 38\u001b[0;31m                 \u001b[0mxq\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mconverted_call\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtf\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcond\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mauto_alpha\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mauto_alpha_computation\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mstatic_alpha_computation\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfscope\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     39\u001b[0m                 \u001b[0;32mtry\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     40\u001b[0m                     \u001b[0mdo_return\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;32mTrue\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n","\u001b[0;32m/content/gdrive/My Drive/matt_schoenbauer_qkeras_project/alt_implementations.py\u001b[0m in \u001b[0;36mauto_alpha_computation\u001b[0;34m()\u001b[0m\n\u001b[1;32m     31\u001b[0m                         \u001b[0;32mtry\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     32\u001b[0m                             \u001b[0mdo_return_2\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;32mTrue\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 33\u001b[0;31m                             \u001b[0mretval__2\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mconverted_call\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_auto_alpha_computation\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfscope_2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     34\u001b[0m                         \u001b[0;32mexcept\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     35\u001b[0m                             \u001b[0mdo_return_2\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;32mFalse\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n","\u001b[0;32m/content/gdrive/My Drive/matt_schoenbauer_qkeras_project/alt_implementations.py\u001b[0m in \u001b[0;36mtf___auto_alpha_computation\u001b[0;34m(self, x)\u001b[0m\n\u001b[1;32m     46\u001b[0m                             \u001b[0;32mraise\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     47\u001b[0m                         \u001b[0;32mreturn\u001b[0m \u001b[0mfscope_2\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mret\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mretval__2\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdo_return_2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 48\u001b[0;31m                 \u001b[0mxq\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mconverted_call\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtf\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcase\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mconverted_call\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtf\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mequal\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0malpha\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mAUTO_ALPHA_ENUM\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfscope\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mautoscale\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mconverted_call\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtf\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mequal\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0malpha\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mAUTO_PO2_ALPHA_ENUM\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfscope\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mpo2_autoscale\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0...\n\u001b[0m\u001b[1;32m     49\u001b[0m                 \u001b[0;32mtry\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     50\u001b[0m                     \u001b[0mdo_return\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;32mTrue\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n","\u001b[0;32m/content/gdrive/My Drive/matt_schoenbauer_qkeras_project/alt_implementations.py\u001b[0m in \u001b[0;36mpo2_autoscale\u001b[0;34m()\u001b[0m\n\u001b[1;32m     36\u001b[0m                         \u001b[0mdo_return_2\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;32mFalse\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     37\u001b[0m                         \u001b[0mretval__2\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mUndefinedReturnValue\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 38\u001b[0;31m                         \u001b[0;34m(\u001b[0m\u001b[0m_alpha_scale\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mint_xq\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mconverted_call\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_po2_autoscale\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0malpha_scale\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfscope_2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     39\u001b[0m                         \u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mconverted_call\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtf\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mprint\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0;34m'po2 alpha_scale'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0m_alpha_scale\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfscope_2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     40\u001b[0m                         \u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mscale\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0m_alpha_scale\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m/\u001b[0m \u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0minteger_repr_scale\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n","\u001b[0;32m/content/gdrive/My Drive/matt_schoenbauer_qkeras_project/alt_implementations.py\u001b[0m in \u001b[0;36mtf___po2_autoscale\u001b[0;34m(self, x, alpha_scale)\u001b[0m\n\u001b[1;32m     46\u001b[0m                         \u001b[0;32mreturn\u001b[0m \u001b[0mfscope_2\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mret\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mretval__2\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdo_return_2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     47\u001b[0m                 \u001b[0mdummy_alpha_scale\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m-\u001b[0m\u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mconverted_call\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtf\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mones_like\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0malpha_scale\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfscope\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 48\u001b[0;31m                 \u001b[0;34m(\u001b[0m\u001b[0m_\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0malpha_scale\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mint_xq\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0m__\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mconverted_call\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtf\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mwhile_loop\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mloop_cond\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mloop_body\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdummy_alpha_scale\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0malpha_scale\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m0\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfscope\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     49\u001b[0m                 \u001b[0;32mtry\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     50\u001b[0m                     \u001b[0mdo_return\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;32mTrue\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n","\u001b[0;31mValueError\u001b[0m: in user code:\n\n    File \"/content/gdrive/My Drive/matt_schoenbauer_qkeras_project/alt_implementations.py\", line 323, in __call__  *\n        xq = tf.cond(\n    File \"/content/gdrive/My Drive/matt_schoenbauer_qkeras_project/alt_implementations.py\", line 339, in auto_alpha_computation  *\n        return self._auto_alpha_computation(x)\n    File \"/content/gdrive/My Drive/matt_schoenbauer_qkeras_project/alt_implementations.py\", line 364, in po2_autoscale  *\n        _alpha_scale, int_xq = self._po2_autoscale(x, alpha_scale)\n    File \"/content/gdrive/My Drive/matt_schoenbauer_qkeras_project/alt_implementations.py\", line 482, in _po2_autoscale  *\n        _, alpha_scale, int_xq, __ = tf.while_loop(\n\n    ValueError: Input tensor `x:0` enters the loop with shape (1,), but has shape (None,) after one iteration. To allow the shape to vary across iterations, use the `shape_invariants` argument of tf.while_loop to specify a less-specific shape.\n"]}]},{"cell_type":"code","source":["import tensorflow as tf\n","\n","class MyClass:\n","\n","  def __init__(self, *, pred):\n","    self.pred = tf.cast(pred, tf.bool)\n","    self.side_effect = tf.Variable(False, dtype=tf.bool)\n","\n","  @tf.function\n","  def __call__(self):\n","      \"\"\"Core quantization function\"\"\"\n","\n","      # Data type conversion\n","\n","      self.side_effect.assign(tf.cond(self.pred, self._true_func, self._false_func))\n","\n","      tf.print(f'{self.side_effect = }')\n","\n","  def _true_func(self):\n","    tf.print('Executing _true_func')\n","    return True\n","\n","  def _false_func(self):\n","    tf.print('Executing _false_func')\n","    return False\n","\n","pred = False\n","c = MyClass(pred=pred)\n","c()\n","print(f'{pred = }, {c.side_effect.numpy() = }\\n')\n"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"MT0AMi3QTIF7","executionInfo":{"status":"ok","timestamp":1679941238662,"user_tz":240,"elapsed":184,"user":{"displayName":"Matt Schoenbauer","userId":"02579049095237637221"}},"outputId":"b24e7c38-f88e-41e5-c2c5-d264014ff454"},"execution_count":3,"outputs":[{"output_type":"stream","name":"stdout","text":["Executing _false_func\n","self.side_effect = <tf.Variable 'Variable:0' shape=() dtype=bool>\n","pred = False, c.side_effect.numpy() = False\n","\n"]}]},{"cell_type":"code","source":["import tensorflow as tf\n","\n","class MyClass:\n","\n","    def __init__(self, *, pred):\n","        self.pred = tf.cast(pred, tf.bool)\n","        self.side_effect = None\n","\n","    @tf.function\n","    def __call__(self):\n","        \"\"\"Core quantization function\"\"\"\n","\n","        result = tf.cond(self.pred, self._true_func, self._false_func)\n","        tf.print(f'{result = }')\n","        return result\n","\n","    def _true_func(self):\n","        tf.print('Executing _true_func')\n","        return tf.constant(True)\n","\n","    def _false_func(self):\n","        tf.print('Executing _false_func')\n","        return tf.constant(False)\n","\n","pred = True\n","c = MyClass(pred=pred)\n","c.side_effect = c()\n","print(f'{pred = }, {c.side_effect = }\\n')\n"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"7G6NhbYkXnQP","executionInfo":{"status":"ok","timestamp":1679942743357,"user_tz":240,"elapsed":164,"user":{"displayName":"Matt Schoenbauer","userId":"02579049095237637221"}},"outputId":"2cb0bfe5-990d-4bbe-8fc5-43c25f3c47b3"},"execution_count":8,"outputs":[{"output_type":"stream","name":"stdout","text":["Executing _true_func\n","result = <tf.Tensor 'cond/Identity:0' shape=() dtype=bool>\n","pred = True, c.side_effect = <tf.Tensor: shape=(), dtype=bool, numpy=True>\n","\n"]}]},{"cell_type":"code","source":["c.side_effect"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"6ua-aRDTXqcb","executionInfo":{"status":"ok","timestamp":1679942398618,"user_tz":240,"elapsed":127,"user":{"displayName":"Matt Schoenbauer","userId":"02579049095237637221"}},"outputId":"32fc34b9-6344-4034-ebcb-54cf781cda97"},"execution_count":5,"outputs":[{"output_type":"execute_result","data":{"text/plain":["<tf.Tensor 'cond/Identity:0' shape=() dtype=bool>"]},"metadata":{},"execution_count":5}]},{"cell_type":"code","source":["tf.convert_to_tensor(True, tf.bool)"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"TC8l_BWc8pbU","executionInfo":{"status":"ok","timestamp":1679942650424,"user_tz":240,"elapsed":4,"user":{"displayName":"Matt Schoenbauer","userId":"02579049095237637221"}},"outputId":"ad741eac-5f26-455d-a35a-618b27c094ed"},"execution_count":7,"outputs":[{"output_type":"execute_result","data":{"text/plain":["<tf.Tensor: shape=(), dtype=bool, numpy=True>"]},"metadata":{},"execution_count":7}]},{"cell_type":"code","execution_count":null,"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"executionInfo":{"elapsed":189,"status":"ok","timestamp":1679883665505,"user":{"displayName":"Matt Schoenbauer","userId":"02579049095237637221"},"user_tz":240},"id":"uUKFOSZdQN0q","outputId":"eedd403c-cfc5-49a5-87a9-e800c9e123e6"},"outputs":[{"output_type":"execute_result","data":{"text/plain":["<tf.Tensor: shape=(), dtype=bool, numpy=True>"]},"metadata":{},"execution_count":124}],"source":["stupid.auto_alpha"]},{"cell_type":"code","execution_count":null,"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"executionInfo":{"elapsed":2,"status":"ok","timestamp":1679792774477,"user":{"displayName":"Matt Schoenbauer","userId":"02579049095237637221"},"user_tz":240},"id":"3DV_C-jJVys4","outputId":"af478001-6545-484f-cb0e-823118c33a02"},"outputs":[{"data":{"text/plain":["<tf.Tensor: shape=(10,), dtype=float32, numpy=\n","array([-0.3828125, -0.0546875, -0.6328125,  0.4609375, -0.9921875,\n","       -0.9921875, -0.4921875,  0.8828125, -0.1015625, -0.6484375],\n","      dtype=float32)>"]},"execution_count":278,"metadata":{},"output_type":"execute_result"}],"source":["alt._auto_alpha_computation(x)"]},{"cell_type":"markdown","metadata":{"id":"iHtCahkEOrNT"},"source":["# **Time Display**"]},{"cell_type":"code","execution_count":103,"metadata":{"id":"T2tmKYS3MBVI","executionInfo":{"status":"ok","timestamp":1679934717415,"user_tz":240,"elapsed":1130,"user":{"displayName":"Matt Schoenbauer","userId":"02579049095237637221"}}},"outputs":[],"source":["def display_runtime_results(quantized_bits_kwargs, func, error_bars=False, field='avg_time'):\n","\n","  data_shapes = (\n","      (1, ), (10, ), (10, 10), (100, 10), (100, 100), (1000, 100),\n","      (1000, 1000), (10000, 1000), (10000, 3000), (10000, 10000)\n","  )\n","  max_iterations = 100\n","  max_time=10\n","  plot=True\n","  hardware = '/GPU:0'\n","  # hardware = '/TPU:0'\n","\n","\n","  print(f'\\n{quantized_bits_kwargs = }\\n')\n","  # Make kwargs hashable\n","  quantized_bits_kwargs = tuple(quantized_bits_kwargs.items())\n","\n","  timing_utils.quantized_bits_speed_tests(func, hardware=hardware, data_shapes=data_shapes,\n","        max_iterations=max_iterations, max_time=max_time, plot=plot, \n","        quantized_bits_kwargs=quantized_bits_kwargs, error_bars=error_bars, field=field)\n","  "]},{"cell_type":"code","execution_count":104,"metadata":{"id":"5aPDO-y1gxQt","executionInfo":{"status":"ok","timestamp":1679934717416,"user_tz":240,"elapsed":3,"user":{"displayName":"Matt Schoenbauer","userId":"02579049095237637221"}}},"outputs":[],"source":["quantized_bits_kwargs_list = [\n","    {},\n","    {'bits': 1},\n","    {'alpha': 'auto'},\n","    {'alpha': 'auto_po2'},    \n","]\n"]},{"cell_type":"code","execution_count":118,"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"executionInfo":{"elapsed":1279,"status":"ok","timestamp":1679935183741,"user":{"displayName":"Matt Schoenbauer","userId":"02579049095237637221"},"user_tz":240},"id":"KqCRc3W-op1v","outputId":"e4b74e2d-dbe5-4dd1-90e6-3ede4aa7cfe8"},"outputs":[{"output_type":"stream","name":"stdout","text":["File alt_implementations.py Updated. Last modified 0:00:35 seconds ago\n"]},{"output_type":"execute_result","data":{"text/plain":["1"]},"metadata":{},"execution_count":118}],"source":["1"]},{"cell_type":"code","execution_count":119,"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":704},"executionInfo":{"elapsed":13001,"status":"error","timestamp":1679935201459,"user":{"displayName":"Matt Schoenbauer","userId":"02579049095237637221"},"user_tz":240},"id":"pl1tGjaE0WQr","outputId":"3d892c71-9655-48f4-95e7-d5cc8d13318e"},"outputs":[{"output_type":"stream","name":"stdout","text":["\n","quantized_bits_kwargs = {}\n","\n"]},{"output_type":"error","ename":"ValueError","evalue":"ignored","traceback":["\u001b[0;31m---------------------------------------------------------------------------\u001b[0m","\u001b[0;31mValueError\u001b[0m                                Traceback (most recent call last)","\u001b[0;32m<ipython-input-119-7f5f245dc1ec>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m      1\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mquantized_bits_kwargs\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mquantized_bits_kwargs_list\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 2\u001b[0;31m   display_runtime_results(\n\u001b[0m\u001b[1;32m      3\u001b[0m       \u001b[0mquantized_bits_kwargs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0malt_implementations\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mquantized_bits2\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0merror_bars\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mFalse\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      4\u001b[0m       field='median_time')\n","\u001b[0;32m<ipython-input-103-66993d13ad9e>\u001b[0m in \u001b[0;36mdisplay_runtime_results\u001b[0;34m(quantized_bits_kwargs, func, error_bars, field)\u001b[0m\n\u001b[1;32m     16\u001b[0m   \u001b[0mquantized_bits_kwargs\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mtuple\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mquantized_bits_kwargs\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mitems\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     17\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 18\u001b[0;31m   timing_utils.quantized_bits_speed_tests(func, hardware=hardware, data_shapes=data_shapes,\n\u001b[0m\u001b[1;32m     19\u001b[0m         \u001b[0mmax_iterations\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mmax_iterations\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mmax_time\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mmax_time\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mplot\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mplot\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     20\u001b[0m         quantized_bits_kwargs=quantized_bits_kwargs, error_bars=error_bars, field=field)\n","\u001b[0;32m/content/gdrive/My Drive/matt_schoenbauer_qkeras_project/timing_utils.py\u001b[0m in \u001b[0;36mquantized_bits_speed_tests\u001b[0;34m(plot, error_bars, field, *args, **kwargs)\u001b[0m\n\u001b[1;32m    103\u001b[0m ):\n\u001b[1;32m    104\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 105\u001b[0;31m     \u001b[0mbaseline_times\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0malt_times\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0m_quantized_bits_speed_test_data\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    106\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    107\u001b[0m     \u001b[0;32mif\u001b[0m \u001b[0mplot\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n","\u001b[0;32m/content/gdrive/My Drive/matt_schoenbauer_qkeras_project/timing_utils.py\u001b[0m in \u001b[0;36m_quantized_bits_speed_test_data\u001b[0;34m(alt_quantized_bits, hardware, data_shapes, max_iterations, max_time, quantized_bits_kwargs)\u001b[0m\n\u001b[1;32m     94\u001b[0m         \u001b[0mqkeras\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mquantized_bits\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m**\u001b[0m\u001b[0mquantized_bits_kwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     95\u001b[0m     )\n\u001b[0;32m---> 96\u001b[0;31m     \u001b[0malt_times\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mhardware_timeit\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0malt_quantized_bits\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m**\u001b[0m\u001b[0mquantized_bits_kwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     97\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     98\u001b[0m     \u001b[0;32mreturn\u001b[0m \u001b[0mbaseline_times\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0malt_times\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n","\u001b[0;32m/content/gdrive/My Drive/matt_schoenbauer_qkeras_project/timing_utils.py\u001b[0m in \u001b[0;36mhardware_timeit\u001b[0;34m(func, hardware, data_shapes, max_iterations, max_time)\u001b[0m\n\u001b[1;32m     51\u001b[0m                 \u001b[0mx\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mtf\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mrandom\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mnormal\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mshape\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     52\u001b[0m                 \u001b[0mstart_time\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mtime\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 53\u001b[0;31m                 \u001b[0m_\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mfunc\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     54\u001b[0m             \u001b[0mrun_time\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mtime\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m-\u001b[0m \u001b[0mstart_time\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     55\u001b[0m             \u001b[0mtotal_time\u001b[0m \u001b[0;34m+=\u001b[0m \u001b[0mrun_time\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n","\u001b[0;32m/usr/local/lib/python3.9/dist-packages/tensorflow/python/util/traceback_utils.py\u001b[0m in \u001b[0;36merror_handler\u001b[0;34m(*args, **kwargs)\u001b[0m\n\u001b[1;32m    151\u001b[0m     \u001b[0;32mexcept\u001b[0m \u001b[0mException\u001b[0m \u001b[0;32mas\u001b[0m \u001b[0me\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    152\u001b[0m       \u001b[0mfiltered_tb\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0m_process_traceback_frames\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0me\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m__traceback__\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 153\u001b[0;31m       \u001b[0;32mraise\u001b[0m \u001b[0me\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mwith_traceback\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfiltered_tb\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mfrom\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    154\u001b[0m     \u001b[0;32mfinally\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    155\u001b[0m       \u001b[0;32mdel\u001b[0m \u001b[0mfiltered_tb\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n","\u001b[0;32m/content/gdrive/My Drive/matt_schoenbauer_qkeras_project/alt_implementations.py\u001b[0m in \u001b[0;36mtf____call__\u001b[0;34m(self, x)\u001b[0m\n\u001b[1;32m     12\u001b[0m                 \u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mconverted_call\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_build\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfscope\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     13\u001b[0m                 \u001b[0mx\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mconverted_call\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mK\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcast_to_floatx\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfscope\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 14\u001b[0;31m                 \u001b[0mxq\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mconverted_call\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtf\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcond\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0muse_sign_function\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mautograph_artifact\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;32mlambda\u001b[0m \u001b[0;34m:\u001b[0m \u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mconverted_call\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_sign_function\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfscope\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mautograph_artifact\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;32mlambda\u001b[0m \u001b[0;34m:\u001b[0m \u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mconverted_call\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_multi_bit_computation\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfscope\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfscope\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     15\u001b[0m                 \u001b[0;32mtry\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     16\u001b[0m                     \u001b[0mdo_return\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;32mTrue\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n","\u001b[0;32m/content/gdrive/My Drive/matt_schoenbauer_qkeras_project/alt_implementations.py\u001b[0m in \u001b[0;36m<lambda>\u001b[0;34m()\u001b[0m\n\u001b[1;32m     12\u001b[0m                 \u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mconverted_call\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_build\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfscope\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     13\u001b[0m                 \u001b[0mx\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mconverted_call\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mK\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcast_to_floatx\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfscope\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 14\u001b[0;31m                 \u001b[0mxq\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mconverted_call\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtf\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcond\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0muse_sign_function\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mautograph_artifact\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;32mlambda\u001b[0m \u001b[0;34m:\u001b[0m \u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mconverted_call\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_sign_function\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfscope\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mautograph_artifact\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;32mlambda\u001b[0m \u001b[0;34m:\u001b[0m \u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mconverted_call\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_multi_bit_computation\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfscope\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfscope\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     15\u001b[0m                 \u001b[0;32mtry\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     16\u001b[0m                     \u001b[0mdo_return\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;32mTrue\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n","\u001b[0;32m/content/gdrive/My Drive/matt_schoenbauer_qkeras_project/alt_implementations.py\u001b[0m in \u001b[0;36mtf___multi_bit_computation\u001b[0;34m(self, x)\u001b[0m\n\u001b[1;32m     36\u001b[0m                             \u001b[0;32mraise\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     37\u001b[0m                         \u001b[0;32mreturn\u001b[0m \u001b[0mfscope_2\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mret\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mretval__2\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdo_return_2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 38\u001b[0;31m                 \u001b[0mxq\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mconverted_call\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtf\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcond\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mauto_alpha\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mauto_alpha_computation\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mstatic_alpha_computation\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfscope\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     39\u001b[0m                 \u001b[0;32mtry\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     40\u001b[0m                     \u001b[0mdo_return\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;32mTrue\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n","\u001b[0;32m/content/gdrive/My Drive/matt_schoenbauer_qkeras_project/alt_implementations.py\u001b[0m in \u001b[0;36mauto_alpha_computation\u001b[0;34m()\u001b[0m\n\u001b[1;32m     31\u001b[0m                         \u001b[0;32mtry\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     32\u001b[0m                             \u001b[0mdo_return_2\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;32mTrue\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 33\u001b[0;31m                             \u001b[0mretval__2\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mconverted_call\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_auto_alpha_computation\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfscope_2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     34\u001b[0m                         \u001b[0;32mexcept\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     35\u001b[0m                             \u001b[0mdo_return_2\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;32mFalse\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n","\u001b[0;32m/content/gdrive/My Drive/matt_schoenbauer_qkeras_project/alt_implementations.py\u001b[0m in \u001b[0;36mtf___auto_alpha_computation\u001b[0;34m(self, x)\u001b[0m\n\u001b[1;32m     46\u001b[0m                             \u001b[0;32mraise\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     47\u001b[0m                         \u001b[0;32mreturn\u001b[0m \u001b[0mfscope_2\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mret\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mretval__2\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdo_return_2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 48\u001b[0;31m                 \u001b[0mxq\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mconverted_call\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtf\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcase\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mconverted_call\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtf\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mequal\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0malpha\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mAUTO_ALPHA_ENUM\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfscope\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mautoscale\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mconverted_call\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtf\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mequal\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0malpha\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mAUTO_PO2_ALPHA_ENUM\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfscope\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mpo2_autoscale\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0...\n\u001b[0m\u001b[1;32m     49\u001b[0m                 \u001b[0;32mtry\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     50\u001b[0m                     \u001b[0mdo_return\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;32mTrue\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n","\u001b[0;32m/content/gdrive/My Drive/matt_schoenbauer_qkeras_project/alt_implementations.py\u001b[0m in \u001b[0;36mpo2_autoscale\u001b[0;34m()\u001b[0m\n\u001b[1;32m     36\u001b[0m                         \u001b[0mdo_return_2\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;32mFalse\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     37\u001b[0m                         \u001b[0mretval__2\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mUndefinedReturnValue\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 38\u001b[0;31m                         \u001b[0;34m(\u001b[0m\u001b[0m_alpha_scale\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mint_xq\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mconverted_call\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_po2_autoscale\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0malpha_scale\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfscope_2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     39\u001b[0m                         \u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mconverted_call\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtf\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mprint\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0;34m'po2 alpha_scale'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0m_alpha_scale\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfscope_2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     40\u001b[0m                         \u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mscale\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0m_alpha_scale\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m/\u001b[0m \u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0minteger_repr_scale\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n","\u001b[0;32m/content/gdrive/My Drive/matt_schoenbauer_qkeras_project/alt_implementations.py\u001b[0m in \u001b[0;36mtf___po2_autoscale\u001b[0;34m(self, x, alpha_scale)\u001b[0m\n\u001b[1;32m     46\u001b[0m                         \u001b[0;32mreturn\u001b[0m \u001b[0mfscope_2\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mret\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mretval__2\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdo_return_2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     47\u001b[0m                 \u001b[0mdummy_alpha_scale\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m-\u001b[0m\u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mconverted_call\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtf\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mones_like\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0malpha_scale\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfscope\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 48\u001b[0;31m                 \u001b[0;34m(\u001b[0m\u001b[0m_\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0malpha_scale\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mint_xq\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0m__\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mconverted_call\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtf\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mwhile_loop\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mloop_cond\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mloop_body\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdummy_alpha_scale\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0malpha_scale\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mag__\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mld\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m0\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mfscope\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     49\u001b[0m                 \u001b[0;32mtry\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     50\u001b[0m                     \u001b[0mdo_return\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;32mTrue\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n","\u001b[0;31mValueError\u001b[0m: in user code:\n\n    File \"/content/gdrive/My Drive/matt_schoenbauer_qkeras_project/alt_implementations.py\", line 323, in __call__  *\n        xq = tf.cond(\n    File \"/content/gdrive/My Drive/matt_schoenbauer_qkeras_project/alt_implementations.py\", line 339, in auto_alpha_computation  *\n        return self._auto_alpha_computation(x)\n    File \"/content/gdrive/My Drive/matt_schoenbauer_qkeras_project/alt_implementations.py\", line 364, in po2_autoscale  *\n        _alpha_scale, int_xq = self._po2_autoscale(x, alpha_scale)\n    File \"/content/gdrive/My Drive/matt_schoenbauer_qkeras_project/alt_implementations.py\", line 482, in _po2_autoscale  *\n        _, alpha_scale, int_xq, __ = tf.while_loop(\n\n    ValueError: Input tensor `x:0` enters the loop with shape (1,), but has shape (None,) after one iteration. To allow the shape to vary across iterations, use the `shape_invariants` argument of tf.while_loop to specify a less-specific shape.\n"]}],"source":["for quantized_bits_kwargs in quantized_bits_kwargs_list:  \n","  display_runtime_results(\n","      quantized_bits_kwargs, alt_implementations.quantized_bits2, error_bars=False, \n","      field='median_time')"]},{"cell_type":"code","source":[],"metadata":{"id":"EboqL4xs773X"},"execution_count":null,"outputs":[]},{"cell_type":"markdown","metadata":{"id":"iysCqbffwiFM"},"source":["# Zero-bit and negative-bit weirdness"]},{"cell_type":"code","execution_count":null,"metadata":{"id":"QKQdvSbxvzEB"},"outputs":[],"source":["q_new = alt_implementations.quantized_bits2(bits=1)\n","q_old = correctness_utils.quantized_bits(bits=-10)\n","q_old_pos = correctness_utils.quantized_bits(bits=0, keep_negative=False)"]},{"cell_type":"code","execution_count":null,"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"executionInfo":{"elapsed":101,"status":"ok","timestamp":1679580352549,"user":{"displayName":"Matt Schoenbauer","userId":"02579049095237637221"},"user_tz":240},"id":"Xe6tLmWTv-gG","outputId":"91c7909e-e155-4ffb-f44d-055a09968e8b"},"outputs":[{"name":"stdout","output_type":"stream","text":["-1.0 1.0\n","0.0 1.0\n"]}],"source":["print(q_old(-1).numpy(), q_old(1).numpy())\n","print(q_old_pos(-1).numpy(), q_old_pos(1).numpy())"]},{"cell_type":"markdown","metadata":{"id":"XDlMgrNzNHjM"},"source":["# Archive"]},{"cell_type":"code","execution_count":null,"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"executionInfo":{"elapsed":173,"status":"ok","timestamp":1679328898933,"user":{"displayName":"Matt Schoenbauer","userId":"02579049095237637221"},"user_tz":240},"id":"x53EYTYHyb2i","outputId":"892db788-d339-4f04-e941-50d7167f311a"},"outputs":[{"data":{"text/plain":["(    iterations  avg_time  time_std           shape       size\n"," 0          100  0.002471  0.001176            (1,)          1\n"," 1          100  0.002088  0.000160           (10,)         10\n"," 2          100  0.002236  0.000691        (10, 10)        100\n"," 3          100  0.002116  0.000342       (100, 10)       1000\n"," 4          100  0.002044  0.000141      (100, 100)      10000\n"," 5          100  0.002122  0.000564     (1000, 100)     100000\n"," 6          100  0.002073  0.000673    (1000, 1000)    1000000\n"," 7          100  0.005878  0.000464   (10000, 1000)   10000000\n"," 8          100  0.012262  0.000605   (10000, 2000)   20000000\n"," 9          100  0.031251  0.001046   (10000, 5000)   50000000\n"," 10         100  0.063647  0.001364  (10000, 10000)  100000000,\n","     iterations  avg_time  time_std           shape       size\n"," 0          100  0.000961  0.003858            (1,)          1\n"," 1          100  0.000897  0.003691           (10,)         10\n"," 2          100  0.000918  0.003761        (10, 10)        100\n"," 3          100  0.001063  0.004938       (100, 10)       1000\n"," 4          100  0.000916  0.003836      (100, 100)      10000\n"," 5          100  0.000903  0.003706     (1000, 100)     100000\n"," 6          100  0.000962  0.003961    (1000, 1000)    1000000\n"," 7          100  0.001087  0.005332   (10000, 1000)   10000000\n"," 8          100  0.001160  0.006344   (10000, 2000)   20000000\n"," 9          100  0.003761  0.031881   (10000, 5000)   50000000\n"," 10         100  0.011218  0.106155  (10000, 10000)  100000000)"]},"execution_count":28,"metadata":{},"output_type":"execute_result"}],"source":["timing_utils._quantized_bits_speed_test_data(quantized_bits2, hardware=hardware, data_shapes=data_shapes,\n","      max_iterations=max_iterations, max_time=max_time)"]},{"cell_type":"code","execution_count":null,"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"executionInfo":{"elapsed":420,"status":"ok","timestamp":1679330469662,"user":{"displayName":"Matt Schoenbauer","userId":"02579049095237637221"},"user_tz":240},"id":"kk6yQrE34YCx","outputId":"430a0bd2-a812-4951-9fd4-042727fd3a89"},"outputs":[{"name":"stdout","output_type":"stream","text":["File alt_implementations.py Updated. Last modified 0:00:50 seconds ago\n"]},{"data":{"text/plain":["False"]},"execution_count":10,"metadata":{},"output_type":"execute_result"}],"source":["bool(tf.keras.backend.learning_phase())"]},{"cell_type":"code","execution_count":null,"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":533},"executionInfo":{"elapsed":578,"status":"ok","timestamp":1679261883151,"user":{"displayName":"Matt Schoenbauer","userId":"02579049095237637221"},"user_tz":240},"id":"daQfmXoDbQF2","outputId":"2fb3d220-f30e-4698-b136-d98f4119bacd"},"outputs":[{"data":{"image/png":"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\n","text/plain":["<Figure size 432x288 with 1 Axes>"]},"metadata":{},"output_type":"display_data"},{"data":{"image/png":"iVBORw0KGgoAAAANSUhEUgAAAXQAAAD4CAYAAAD8Zh1EAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjcuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/bCgiHAAAACXBIWXMAAAsTAAALEwEAmpwYAAAQ9klEQVR4nO3df6zdd13H8eeLlkJkwwEtY24tHdIoxV/M64ZAkMAkG5oWAyRbYmRxpBBtRDHRkpnFzEQDJJgYaqTiIhpgjCla9JLxO2rMZu9wP+jGoNRh20xWfjhcCBtlb/+453aHu3t7z13P93Pu+fb5SG76/fHpOe9+e/vq937e3+/5pqqQJE2/J026AEnSeBjoktQTBrok9YSBLkk9YaBLUk+sn9Qbb9y4sbZu3Tqpt5ekqXTbbbd9vao2LbVvYoG+detW5ubmJvX2kjSVknx1uX1OuUhSTxjoktQTBrok9YSBLkk9YaBLUk+MFOhJLktyb5JDSfYssf+qJMeT3D74etP4S5UkncqKly0mWQfsBX4ROAocSLK/qu5eNPTDVbW7gxolSSMY5Tr0i4FDVXUYIMkNwE5gcaBLGvLIiUf563//Lx767olJl6I15lUvOJef3nzO2F93lEA/HzgytH4UuGSJca9L8nLgS8DvVNWRxQOS7AJ2AWzZsmX11UpT5K5j/8sfz34RgGTCxWhNefbTnzqxQB/Fx4APVdXDSd4MvB945eJBVbUP2AcwMzPjkzXUa99/dP7XD7zpEl76/I2TLUZnhFGaoseAzUPrFwy2nVRV36iqhwer7wN+djzlSdPLp4GptVEC/QCwLcmFSTYAVwD7hwckOW9odQdwz/hKlCSNYsUpl6o6kWQ3cDOwDri+qg4muQ6Yq6r9wG8l2QGcAL4JXNVhzdJUcfpcrYw0h15Vs8Dsom3XDi2/HXj7eEuTppsTLmrNO0UlqScMdKkjJ3uizrmoEQNdknrCQJc6Fk/R1YiBLnWkbIuqMQNdknrCQJc65ue4qBUDXeqKMy5qzECXOuYJulox0KWOeIKu1gx0SeoJA13qWOyKqhEDXeqIH4eu1gx0qWOeoKsVA13qiHeKqjUDXZJ6wkCXOuaMi1ox0KWO2BRVawa61DGbomrFQJc64gm6WjPQJaknDHSpc865qA0DXepI2RVVYwa61DGbomrFQJc64vm5WjPQJaknDHSpY864qBUDXeqKcy5qzECXOuYDLtSKgS5JPWGgSx3x89DV2kiBnuSyJPcmOZRkzynGvS5JJZkZX4nSdHPCRa2sGOhJ1gF7gcuB7cCVSbYvMe5s4K3AreMuUppG3iiq1kY5Q78YOFRVh6vqEeAGYOcS4/4IeAfw3THWJ009e6JqZZRAPx84MrR+dLDtpCQXAZur6p9P9UJJdiWZSzJ3/PjxVRcrSVreaTdFkzwJeDfwuyuNrap9VTVTVTObNm063beW1jSnXNTaKIF+DNg8tH7BYNuCs4GfAD6X5D7gxcB+G6PSvNgWVSOjBPoBYFuSC5NsAK4A9i/srKoHq2pjVW2tqq3ALcCOqprrpGJpSniCrtZWDPSqOgHsBm4G7gFurKqDSa5LsqPrAqVpZ1NUrawfZVBVzQKzi7Zdu8zYV5x+WZKk1fJOUakjPrFIrRnoktQTBrrUEc/P1ZqBLnXMpqhaMdAlqScMdKkj9kTVmoEudcw7RdWKgS51xlN0tWWgS1JPGOhSx7zKRa0Y6FJHbIqqNQNd6phn6GrFQJeknjDQpY4446LWDHSpY16HrlYMdKkjNkXVmoEudcymqFox0CWpJwx0qSNlW1SNGehSx5xxUSsGutQRm6JqzUCXOmZTVK0Y6JLUEwa61BFnXNSagS51zjkXtWGgSx0pu6JqzECXOmZTVK0Y6JLUEwa6JPWEgS51zBkXtTJSoCe5LMm9SQ4l2bPE/rckuSvJ7Un+Lcn28ZcqTRd7omptxUBPsg7YC1wObAeuXCKwP1hVP1lVPwO8E3j3uAuVplXsiqqRUc7QLwYOVdXhqnoEuAHYOTygqr49tPo0vKdCkppbP8KY84EjQ+tHgUsWD0rym8DbgA3AK5d6oSS7gF0AW7ZsWW2t0lTx43PV2tiaolW1t6p+FPh94A+WGbOvqmaqambTpk3jemtpTXPCRa2MEujHgM1D6xcMti3nBuC1p1GTJOkJGCXQDwDbklyYZANwBbB/eECSbUOrvwR8eXwlStNp4SoXe6JqZcU59Ko6kWQ3cDOwDri+qg4muQ6Yq6r9wO4klwLfA74FvLHLoiVJjzdKU5SqmgVmF227dmj5rWOuS5p6Xoeu1rxTVOpYbIuqEQNdknrCQJc6sjDjYlNUrRjoktQTBrrUEZ9YpNYMdEnqCQNdknrCQJc6YlNUrRnoktQTBrrUFXuiasxAlzrmE4vUioEuST1hoEsdWXhikefnasVAl6SeMNCljnijqFoz0KWO2RNVKwa6JPWEgS515OSdorZF1YiBLkk9YaBLUk8Y6FJHFq5ysSmqVgx0SeoJA13qSPnpXGrMQJc65oyLWjHQJaknDHSpI/XYhehSEwa6JPWEgS51xJaoWjPQpY55679aMdAlqScMdKkrg66od4qqlZECPcllSe5NcijJniX2vy3J3UnuTPLpJM8df6mSpFNZMdCTrAP2ApcD24Erk2xfNOw/gZmq+ingJuCd4y5UmjY2RdXaKGfoFwOHqupwVT0C3ADsHB5QVZ+tqu8MVm8BLhhvmdL0csZFrYwS6OcDR4bWjw62Ledq4ONL7UiyK8lckrnjx4+PXqUkaUVjbYom+VVgBnjXUvural9VzVTVzKZNm8b51tKa89jH53qOrjbWjzDmGLB5aP2CwbYfkORS4BrgF6rq4fGUJ0ka1Shn6AeAbUkuTLIBuALYPzwgyYuA9wI7quqB8ZcpSVrJioFeVSeA3cDNwD3AjVV1MMl1SXYMhr0LOAv4SJLbk+xf5uWkM0YtXIc+4Tp05hhlyoWqmgVmF227dmj50jHXJUlaJe8UlTpy8tNzPUVXIwa6JPWEgS5JPWGgSx05eR26bVE1YqBLUk8Y6FJHTn44lyfoasRAl6SeMNAlqScMdKkj5ROL1JiBLkk9YaBLHfMEXa0Y6JLUEwa6JPWEgS51xCcWqTUDXZJ6wkCXOlL4gAu1ZaBLUk8Y6JLUEwa61JHHmqKTrUNnDgNdknrCQJc65gMu1IqBLnWkVh4ijZWBLkk9YaBLHbEpqtYMdEnqCQNdknrCQJc6UrZF1ZiBLkk9YaBLHbEpqtYMdEnqiZECPcllSe5NcijJniX2vzzJ55OcSPL68ZcpTS/vFFUrKwZ6knXAXuByYDtwZZLti4b9N3AV8MFxFyhJGs36EcZcDByqqsMASW4AdgJ3LwyoqvsG+x7toEZJ0ghGmXI5HzgytH50sG3VkuxKMpdk7vjx40/kJaSpUYOuqE1RtdK0KVpV+6pqpqpmNm3a1PKtJan3Rgn0Y8DmofULBtskSWvIKIF+ANiW5MIkG4ArgP3dliVNv5PXoU+2DJ1BVgz0qjoB7AZuBu4Bbqyqg0muS7IDIMnPJTkKvAF4b5KDXRYtSXq8Ua5yoapmgdlF264dWj7A/FSMpIGFT3KJXVE14p2iktQTBrok9YSBLnXEpqhaM9AlqScMdKlj9kTVioEudcQnFqk1A12SesJAlzry2BOLnHNRGwa6JPWEgS5JPWGgSx2xJarWDHRJ6gkDXepKldegqykDXZJ6wkCXOuQJuloy0KWO2BRVawa6JPWEgS51pMq7RNWWgS5JPWGgSx3y/FwtGehSR/z4XLVmoEtSTxjoUkfmm6KTrkJnEgNdknrCQJc6FNuiashAlzpiS1StGeiS1BMGutQlZ1zUkIEudaScc1FjBrrUIU/Q1ZKBLnXEO0XV2kiBnuSyJPcmOZRkzxL7n5Lkw4P9tybZOvZKJUmntGKgJ1kH7AUuB7YDVybZvmjY1cC3qur5wJ8C7xh3odI08k5RtbR+hDEXA4eq6jBAkhuAncDdQ2N2An84WL4JeE+SVI2/LXTjgSP85b8eHvfLSmN3/KGHJ12CzjCjBPr5wJGh9aPAJcuNqaoTSR4EngV8fXhQkl3ALoAtW7Y8oYLP+aEns+3cs57Q75Va2nbuWWw/7+mTLkNnkFECfWyqah+wD2BmZuYJnb2/+oXP4dUvfM5Y65KkPhilKXoM2Dy0fsFg25JjkqwHfhj4xjgKlCSNZpRAPwBsS3Jhkg3AFcD+RWP2A28cLL8e+EwX8+eSpOWtOOUymBPfDdwMrAOur6qDSa4D5qpqP/BXwN8mOQR8k/nQlyQ1NNIcelXNArOLtl07tPxd4A3jLU2StBreKSpJPWGgS1JPGOiS1BMGuiT1RCZ1dWGS48BXn+Bv38iiu1DXCOtaHetavbVam3WtzunU9dyq2rTUjokF+ulIMldVM5OuYzHrWh3rWr21Wpt1rU5XdTnlIkk9YaBLUk9Ma6Dvm3QBy7Cu1bGu1VurtVnX6nRS11TOoUuSHm9az9AlSYsY6JLUE1MR6EneleSLSe5M8tEk5ywz7pQPs+6grjckOZjk0STLXoKU5L4kdyW5PcncGqqr9fF6ZpJPJvny4NdnLDPu+4NjdXuSxR/VPM561uTDz0eo66okx4eO0Zsa1XV9kgeSfGGZ/UnyZ4O670xy0Rqp6xVJHhw6XtcuNW7MNW1O8tkkdw/+Lb51iTHjP15Vtea/gFcD6wfL7wDescSYdcBXgOcBG4A7gO0d1/UC4MeAzwEzpxh3H7Cx4fFasa4JHa93AnsGy3uW+nsc7HuowTFa8c8P/AbwF4PlK4APr5G6rgLe0+r7aeh9Xw5cBHxhmf2vAT4OBHgxcOsaqesVwD81PlbnARcNls8GvrTE3+PYj9dUnKFX1Seq6sRg9Rbmn5q02MmHWVfVI8DCw6y7rOueqrq3y/d4Ikasq/nxGrz++wfL7wde2/H7ncoof/7hem8CXpUka6Cuiaiqf2H+eQfL2Qn8Tc27BTgnyXlroK7mqur+qvr8YPn/gHuYf/bysLEfr6kI9EV+nfn/1RZb6mHWiw/gpBTwiSS3DR6UvRZM4nidW1X3D5b/Bzh3mXFPTTKX5JYkr+2ollH+/D/w8HNg4eHnXRr17+V1gx/Tb0qyeYn9k7CW/w3+fJI7knw8yQtbvvFgqu5FwK2Ldo39eDV9SPSpJPkUsNTTn6+pqn8cjLkGOAF8YC3VNYKXVdWxJM8GPpnki4OziknXNXanqmt4paoqyXLXzD53cLyeB3wmyV1V9ZVx1zrFPgZ8qKoeTvJm5n+KeOWEa1rLPs/899RDSV4D/AOwrcUbJzkL+Dvgt6vq212/35oJ9Kq69FT7k1wF/DLwqhpMQC0yysOsx17XiK9xbPDrA0k+yvyP1acV6GOoq/nxSvK1JOdV1f2DHy0fWOY1Fo7X4SSfY/7sZtyBvpqHnx9Nu4efr1hXVQ3X8D7mexNrQSffU6drOEirajbJnyfZWFWdfmhXkiczH+YfqKq/X2LI2I/XVEy5JLkM+D1gR1V9Z5lhozzMurkkT0ty9sIy8w3eJbvxjU3ieA0/TPyNwON+kkjyjCRPGSxvBF4K3N1BLWv14ecr1rVonnUH8/Oza8F+4NcGV2+8GHhwaIptYpI8Z6H3keRi5nOv0/+YB+/3V8A9VfXuZYaN/3i17PyeRsf4EPNzTbcPvhauPPgRYHZR1/hLzJ/NXdOgrl9hft7rYeBrwM2L62L+aoU7Bl8H10pdEzpezwI+DXwZ+BTwzMH2GeB9g+WXAHcNjtddwNUd1vO4Pz9wHfMnDgBPBT4y+P77D+B5XR+jEev6k8H30h3AZ4Efb1TXh4D7ge8Nvr+uBt4CvGWwP8DeQd13cYorvxrXtXvoeN0CvKRBTS9jvnd251Buvabr4+Wt/5LUE1Mx5SJJWpmBLkk9YaBLUk8Y6JLUEwa6JPWEgS5JPWGgS1JP/D8DU1okYU39AwAAAABJRU5ErkJggg==\n","text/plain":["<Figure size 432x288 with 1 Axes>"]},"metadata":{},"output_type":"display_data"}],"source":["def compute_func(function, x_min, x_max, num_points):\n","\n","  x = np.linspace(x_min, x_max, num=num_points)\n","  return x, function(x)\n","\n","\n","def plot_func(function, x_min, x_max, num_points):\n","\n","  x, f_x = compute_func(function, x_min, x_max, num_points)\n","  plt.plot(x, f_x)\n","  plt.show()\n","\n","\n","bits = 1\n","integer = 0\n","symmetric = True\n","keep_negative = False\n","qnoise_factor = 1.0\n","\n","qbits = qkeras.quantized_bits(bits=bits,integer=integer,\n","symmetric=symmetric,\n","keep_negative=keep_negative,\n","qnoise_factor=qnoise_factor,)\n","\n","qbits2 = alt_implementations.quantized_bits2(bits=bits,integer=integer,\n","symmetric=symmetric,\n","keep_negative=keep_negative,\n","qnoise_factor=qnoise_factor,)\n","\n","x_min = -2\n","x_max = 2\n","num_points = 1000\n","plot_func(qbits,  x_min, x_max, num_points)\n","plot_func(qbits2,  x_min, x_max, num_points)"]},{"cell_type":"code","execution_count":null,"metadata":{"id":"VnJqNJZAhZpy"},"outputs":[],"source":[]},{"cell_type":"markdown","metadata":{"id":"zNSQP42vhZ9t"},"source":["# Experiments with tf.cond\n","\n","Conclusion- No domination either way between tf.cond and if/else statements in Python "]},{"cell_type":"code","execution_count":null,"metadata":{"id":"24BdlKKJs_w6"},"outputs":[],"source":["@tf.function\n","def py_cond(x):\n","\n","  if x.shape[0] > tf.constant(10):\n","    return 1\n","  else:\n","    return 0\n","\n","@tf.function\n","def tf_cond(x):\n","\n","  return tf.cond(tf.greater(x.shape[0], 10), lambda: 1, lambda: 0)\n","\n","@tf.function\n","def tf_cond2(x):\n","\n","  return tf.cond(tf.greater(x.shape[0], tf.constant(10)), lambda: 1, lambda: 0)"]},{"cell_type":"code","execution_count":null,"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"executionInfo":{"elapsed":488,"status":"ok","timestamp":1679338409431,"user":{"displayName":"Matt Schoenbauer","userId":"02579049095237637221"},"user_tz":240},"id":"38m__KYGXjlJ","outputId":"e321a71b-26d8-469f-a1ce-b7f2b24bb42e"},"outputs":[{"name":"stdout","output_type":"stream","text":["File timing_utils.py Updated. Last modified 0:00:25 seconds ago\n"]}],"source":["import timing_utils"]},{"cell_type":"code","execution_count":null,"metadata":{"id":"O3-roSbYR-eG"},"outputs":[],"source":["data_shapes = (\n","    (1, ), (10, ), (10, 10), (100, 10), (100, 100), (1000, 100),\n","    (1000, 1000), (10000, 1000), (10000, 2000), (10000, 5000), (10000, 10000)\n",")\n","max_iterations = 200\n","max_time=10\n","plot=True\n","hardware = '/GPU:0'\n","# hardware = '/TPU:0'\n","# hardware = '/CPU:0'\n","bits = 8\n","\n","py_data = timing_utils.hardware_timeit(py_cond, hardware=hardware, data_shapes=data_shapes,\n","      max_iterations=max_iterations, max_time=max_time)\n","tf_data = timing_utils.hardware_timeit(tf_cond, hardware=hardware, data_shapes=data_shapes,\n","      max_iterations=max_iterations, max_time=max_time)\n","tf_data2 = timing_utils.hardware_timeit(tf_cond2, hardware=hardware, data_shapes=data_shapes,\n","      max_iterations=max_iterations, max_time=max_time)"]},{"cell_type":"code","execution_count":null,"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":1000},"executionInfo":{"elapsed":14,"status":"ok","timestamp":1679340953706,"user":{"displayName":"Matt Schoenbauer","userId":"02579049095237637221"},"user_tz":240},"id":"dRQqc3R5TIJH","outputId":"be437a09-e7c9-48af-9ea6-f0a849d4d98b"},"outputs":[{"data":{"text/html":["\n","  <div id=\"df-c5c6b70a-d4d4-4df1-9c7c-31c2721ad09a\">\n","    <div class=\"colab-df-container\">\n","      <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>iterations</th>\n","      <th>avg_time</th>\n","      <th>time_std</th>\n","      <th>shape</th>\n","      <th>size</th>\n","    </tr>\n","  </thead>\n","  <tbody>\n","    <tr>\n","      <th>0</th>\n","      <td>200</td>\n","      <td>0.000677</td>\n","      <td>0.000206</td>\n","      <td>(1,)</td>\n","      <td>1</td>\n","    </tr>\n","    <tr>\n","      <th>1</th>\n","      <td>200</td>\n","      <td>0.000560</td>\n","      <td>0.000154</td>\n","      <td>(10,)</td>\n","      <td>10</td>\n","    </tr>\n","    <tr>\n","      <th>2</th>\n","      <td>200</td>\n","      <td>0.000474</td>\n","      <td>0.000081</td>\n","      <td>(10, 10)</td>\n","      <td>100</td>\n","    </tr>\n","    <tr>\n","      <th>3</th>\n","      <td>200</td>\n","      <td>0.000468</td>\n","      <td>0.000080</td>\n","      <td>(100, 10)</td>\n","      <td>1000</td>\n","    </tr>\n","    <tr>\n","      <th>4</th>\n","      <td>200</td>\n","      <td>0.000479</td>\n","      <td>0.000212</td>\n","      <td>(100, 100)</td>\n","      <td>10000</td>\n","    </tr>\n","    <tr>\n","      <th>5</th>\n","      <td>200</td>\n","      <td>0.000451</td>\n","      <td>0.000026</td>\n","      <td>(1000, 100)</td>\n","      <td>100000</td>\n","    </tr>\n","    <tr>\n","      <th>6</th>\n","      <td>200</td>\n","      <td>0.000455</td>\n","      <td>0.000031</td>\n","      <td>(1000, 1000)</td>\n","      <td>1000000</td>\n","    </tr>\n","    <tr>\n","      <th>7</th>\n","      <td>200</td>\n","      <td>0.000482</td>\n","      <td>0.000132</td>\n","      <td>(10000, 1000)</td>\n","      <td>10000000</td>\n","    </tr>\n","    <tr>\n","      <th>8</th>\n","      <td>200</td>\n","      <td>0.000476</td>\n","      <td>0.000040</td>\n","      <td>(10000, 2000)</td>\n","      <td>20000000</td>\n","    </tr>\n","    <tr>\n","      <th>9</th>\n","      <td>200</td>\n","      <td>0.000521</td>\n","      <td>0.000069</td>\n","      <td>(10000, 5000)</td>\n","      <td>50000000</td>\n","    </tr>\n","    <tr>\n","      <th>10</th>\n","      <td>200</td>\n","      <td>0.000642</td>\n","      <td>0.000168</td>\n","      <td>(10000, 10000)</td>\n","      <td>100000000</td>\n","    </tr>\n","  </tbody>\n","</table>\n","</div>\n","      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-c5c6b70a-d4d4-4df1-9c7c-31c2721ad09a')\"\n","              title=\"Convert this dataframe to an interactive table.\"\n","              style=\"display:none;\">\n","        \n","  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n","       width=\"24px\">\n","    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n","    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n","  </svg>\n","      </button>\n","      \n","  <style>\n","    .colab-df-container {\n","      display:flex;\n","      flex-wrap:wrap;\n","      gap: 12px;\n","    }\n","\n","    .colab-df-convert {\n","      background-color: #E8F0FE;\n","      border: none;\n","      border-radius: 50%;\n","      cursor: pointer;\n","      display: none;\n","      fill: #1967D2;\n","      height: 32px;\n","      padding: 0 0 0 0;\n","      width: 32px;\n","    }\n","\n","    .colab-df-convert:hover {\n","      background-color: #E2EBFA;\n","      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n","      fill: #174EA6;\n","    }\n","\n","    [theme=dark] .colab-df-convert {\n","      background-color: #3B4455;\n","      fill: #D2E3FC;\n","    }\n","\n","    [theme=dark] .colab-df-convert:hover {\n","      background-color: #434B5C;\n","      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n","      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n","      fill: #FFFFFF;\n","    }\n","  </style>\n","\n","      <script>\n","        const buttonEl =\n","          document.querySelector('#df-c5c6b70a-d4d4-4df1-9c7c-31c2721ad09a button.colab-df-convert');\n","        buttonEl.style.display =\n","          google.colab.kernel.accessAllowed ? 'block' : 'none';\n","\n","        async function convertToInteractive(key) {\n","          const element = document.querySelector('#df-c5c6b70a-d4d4-4df1-9c7c-31c2721ad09a');\n","          const dataTable =\n","            await google.colab.kernel.invokeFunction('convertToInteractive',\n","                                                     [key], {});\n","          if (!dataTable) return;\n","\n","          const docLinkHtml = 'Like what you see? Visit the ' +\n","            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n","            + ' to learn more about interactive tables.';\n","          element.innerHTML = '';\n","          dataTable['output_type'] = 'display_data';\n","          await google.colab.output.renderOutput(dataTable, element);\n","          const docLink = document.createElement('div');\n","          docLink.innerHTML = docLinkHtml;\n","          element.appendChild(docLink);\n","        }\n","      </script>\n","    </div>\n","  </div>\n","  "],"text/plain":["    iterations  avg_time  time_std           shape       size\n","0          200  0.000677  0.000206            (1,)          1\n","1          200  0.000560  0.000154           (10,)         10\n","2          200  0.000474  0.000081        (10, 10)        100\n","3          200  0.000468  0.000080       (100, 10)       1000\n","4          200  0.000479  0.000212      (100, 100)      10000\n","5          200  0.000451  0.000026     (1000, 100)     100000\n","6          200  0.000455  0.000031    (1000, 1000)    1000000\n","7          200  0.000482  0.000132   (10000, 1000)   10000000\n","8          200  0.000476  0.000040   (10000, 2000)   20000000\n","9          200  0.000521  0.000069   (10000, 5000)   50000000\n","10         200  0.000642  0.000168  (10000, 10000)  100000000"]},"metadata":{},"output_type":"display_data"},{"data":{"text/html":["\n","  <div id=\"df-511cbe5e-14fc-41ef-8d19-2089c3bf7354\">\n","    <div class=\"colab-df-container\">\n","      <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>iterations</th>\n","      <th>avg_time</th>\n","      <th>time_std</th>\n","      <th>shape</th>\n","      <th>size</th>\n","    </tr>\n","  </thead>\n","  <tbody>\n","    <tr>\n","      <th>0</th>\n","      <td>200</td>\n","      <td>0.000618</td>\n","      <td>0.000071</td>\n","      <td>(1,)</td>\n","      <td>1</td>\n","    </tr>\n","    <tr>\n","      <th>1</th>\n","      <td>200</td>\n","      <td>0.000612</td>\n","      <td>0.000023</td>\n","      <td>(10,)</td>\n","      <td>10</td>\n","    </tr>\n","    <tr>\n","      <th>2</th>\n","      <td>200</td>\n","      <td>0.000616</td>\n","      <td>0.000087</td>\n","      <td>(10, 10)</td>\n","      <td>100</td>\n","    </tr>\n","    <tr>\n","      <th>3</th>\n","      <td>200</td>\n","      <td>0.000644</td>\n","      <td>0.000156</td>\n","      <td>(100, 10)</td>\n","      <td>1000</td>\n","    </tr>\n","    <tr>\n","      <th>4</th>\n","      <td>200</td>\n","      <td>0.000470</td>\n","      <td>0.000059</td>\n","      <td>(100, 100)</td>\n","      <td>10000</td>\n","    </tr>\n","    <tr>\n","      <th>5</th>\n","      <td>200</td>\n","      <td>0.000465</td>\n","      <td>0.000056</td>\n","      <td>(1000, 100)</td>\n","      <td>100000</td>\n","    </tr>\n","    <tr>\n","      <th>6</th>\n","      <td>200</td>\n","      <td>0.000443</td>\n","      <td>0.000037</td>\n","      <td>(1000, 1000)</td>\n","      <td>1000000</td>\n","    </tr>\n","    <tr>\n","      <th>7</th>\n","      <td>200</td>\n","      <td>0.000447</td>\n","      <td>0.000019</td>\n","      <td>(10000, 1000)</td>\n","      <td>10000000</td>\n","    </tr>\n","    <tr>\n","      <th>8</th>\n","      <td>200</td>\n","      <td>0.000506</td>\n","      <td>0.000268</td>\n","      <td>(10000, 2000)</td>\n","      <td>20000000</td>\n","    </tr>\n","    <tr>\n","      <th>9</th>\n","      <td>200</td>\n","      <td>0.000454</td>\n","      <td>0.000049</td>\n","      <td>(10000, 5000)</td>\n","      <td>50000000</td>\n","    </tr>\n","    <tr>\n","      <th>10</th>\n","      <td>200</td>\n","      <td>0.000463</td>\n","      <td>0.000049</td>\n","      <td>(10000, 10000)</td>\n","      <td>100000000</td>\n","    </tr>\n","  </tbody>\n","</table>\n","</div>\n","      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-511cbe5e-14fc-41ef-8d19-2089c3bf7354')\"\n","              title=\"Convert this dataframe to an interactive table.\"\n","              style=\"display:none;\">\n","        \n","  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n","       width=\"24px\">\n","    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n","    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n","  </svg>\n","      </button>\n","      \n","  <style>\n","    .colab-df-container {\n","      display:flex;\n","      flex-wrap:wrap;\n","      gap: 12px;\n","    }\n","\n","    .colab-df-convert {\n","      background-color: #E8F0FE;\n","      border: none;\n","      border-radius: 50%;\n","      cursor: pointer;\n","      display: none;\n","      fill: #1967D2;\n","      height: 32px;\n","      padding: 0 0 0 0;\n","      width: 32px;\n","    }\n","\n","    .colab-df-convert:hover {\n","      background-color: #E2EBFA;\n","      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n","      fill: #174EA6;\n","    }\n","\n","    [theme=dark] .colab-df-convert {\n","      background-color: #3B4455;\n","      fill: #D2E3FC;\n","    }\n","\n","    [theme=dark] .colab-df-convert:hover {\n","      background-color: #434B5C;\n","      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n","      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n","      fill: #FFFFFF;\n","    }\n","  </style>\n","\n","      <script>\n","        const buttonEl =\n","          document.querySelector('#df-511cbe5e-14fc-41ef-8d19-2089c3bf7354 button.colab-df-convert');\n","        buttonEl.style.display =\n","          google.colab.kernel.accessAllowed ? 'block' : 'none';\n","\n","        async function convertToInteractive(key) {\n","          const element = document.querySelector('#df-511cbe5e-14fc-41ef-8d19-2089c3bf7354');\n","          const dataTable =\n","            await google.colab.kernel.invokeFunction('convertToInteractive',\n","                                                     [key], {});\n","          if (!dataTable) return;\n","\n","          const docLinkHtml = 'Like what you see? Visit the ' +\n","            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n","            + ' to learn more about interactive tables.';\n","          element.innerHTML = '';\n","          dataTable['output_type'] = 'display_data';\n","          await google.colab.output.renderOutput(dataTable, element);\n","          const docLink = document.createElement('div');\n","          docLink.innerHTML = docLinkHtml;\n","          element.appendChild(docLink);\n","        }\n","      </script>\n","    </div>\n","  </div>\n","  "],"text/plain":["    iterations  avg_time  time_std           shape       size\n","0          200  0.000618  0.000071            (1,)          1\n","1          200  0.000612  0.000023           (10,)         10\n","2          200  0.000616  0.000087        (10, 10)        100\n","3          200  0.000644  0.000156       (100, 10)       1000\n","4          200  0.000470  0.000059      (100, 100)      10000\n","5          200  0.000465  0.000056     (1000, 100)     100000\n","6          200  0.000443  0.000037    (1000, 1000)    1000000\n","7          200  0.000447  0.000019   (10000, 1000)   10000000\n","8          200  0.000506  0.000268   (10000, 2000)   20000000\n","9          200  0.000454  0.000049   (10000, 5000)   50000000\n","10         200  0.000463  0.000049  (10000, 10000)  100000000"]},"metadata":{},"output_type":"display_data"},{"data":{"text/html":["\n","  <div id=\"df-027aa029-dc5e-4566-bdf3-a1d3a0dc9cd9\">\n","    <div class=\"colab-df-container\">\n","      <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>iterations</th>\n","      <th>avg_time</th>\n","      <th>time_std</th>\n","      <th>shape</th>\n","      <th>size</th>\n","    </tr>\n","  </thead>\n","  <tbody>\n","    <tr>\n","      <th>0</th>\n","      <td>200</td>\n","      <td>0.000459</td>\n","      <td>0.000031</td>\n","      <td>(1,)</td>\n","      <td>1</td>\n","    </tr>\n","    <tr>\n","      <th>1</th>\n","      <td>200</td>\n","      <td>0.000473</td>\n","      <td>0.000065</td>\n","      <td>(10,)</td>\n","      <td>10</td>\n","    </tr>\n","    <tr>\n","      <th>2</th>\n","      <td>200</td>\n","      <td>0.000454</td>\n","      <td>0.000026</td>\n","      <td>(10, 10)</td>\n","      <td>100</td>\n","    </tr>\n","    <tr>\n","      <th>3</th>\n","      <td>200</td>\n","      <td>0.000457</td>\n","      <td>0.000025</td>\n","      <td>(100, 10)</td>\n","      <td>1000</td>\n","    </tr>\n","    <tr>\n","      <th>4</th>\n","      <td>200</td>\n","      <td>0.000471</td>\n","      <td>0.000056</td>\n","      <td>(100, 100)</td>\n","      <td>10000</td>\n","    </tr>\n","    <tr>\n","      <th>5</th>\n","      <td>200</td>\n","      <td>0.000461</td>\n","      <td>0.000045</td>\n","      <td>(1000, 100)</td>\n","      <td>100000</td>\n","    </tr>\n","    <tr>\n","      <th>6</th>\n","      <td>200</td>\n","      <td>0.000458</td>\n","      <td>0.000045</td>\n","      <td>(1000, 1000)</td>\n","      <td>1000000</td>\n","    </tr>\n","    <tr>\n","      <th>7</th>\n","      <td>200</td>\n","      <td>0.000454</td>\n","      <td>0.000026</td>\n","      <td>(10000, 1000)</td>\n","      <td>10000000</td>\n","    </tr>\n","    <tr>\n","      <th>8</th>\n","      <td>200</td>\n","      <td>0.000460</td>\n","      <td>0.000058</td>\n","      <td>(10000, 2000)</td>\n","      <td>20000000</td>\n","    </tr>\n","    <tr>\n","      <th>9</th>\n","      <td>200</td>\n","      <td>0.000464</td>\n","      <td>0.000038</td>\n","      <td>(10000, 5000)</td>\n","      <td>50000000</td>\n","    </tr>\n","    <tr>\n","      <th>10</th>\n","      <td>200</td>\n","      <td>0.000468</td>\n","      <td>0.000046</td>\n","      <td>(10000, 10000)</td>\n","      <td>100000000</td>\n","    </tr>\n","  </tbody>\n","</table>\n","</div>\n","      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-027aa029-dc5e-4566-bdf3-a1d3a0dc9cd9')\"\n","              title=\"Convert this dataframe to an interactive table.\"\n","              style=\"display:none;\">\n","        \n","  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n","       width=\"24px\">\n","    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n","    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n","  </svg>\n","      </button>\n","      \n","  <style>\n","    .colab-df-container {\n","      display:flex;\n","      flex-wrap:wrap;\n","      gap: 12px;\n","    }\n","\n","    .colab-df-convert {\n","      background-color: #E8F0FE;\n","      border: none;\n","      border-radius: 50%;\n","      cursor: pointer;\n","      display: none;\n","      fill: #1967D2;\n","      height: 32px;\n","      padding: 0 0 0 0;\n","      width: 32px;\n","    }\n","\n","    .colab-df-convert:hover {\n","      background-color: #E2EBFA;\n","      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n","      fill: #174EA6;\n","    }\n","\n","    [theme=dark] .colab-df-convert {\n","      background-color: #3B4455;\n","      fill: #D2E3FC;\n","    }\n","\n","    [theme=dark] .colab-df-convert:hover {\n","      background-color: #434B5C;\n","      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n","      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n","      fill: #FFFFFF;\n","    }\n","  </style>\n","\n","      <script>\n","        const buttonEl =\n","          document.querySelector('#df-027aa029-dc5e-4566-bdf3-a1d3a0dc9cd9 button.colab-df-convert');\n","        buttonEl.style.display =\n","          google.colab.kernel.accessAllowed ? 'block' : 'none';\n","\n","        async function convertToInteractive(key) {\n","          const element = document.querySelector('#df-027aa029-dc5e-4566-bdf3-a1d3a0dc9cd9');\n","          const dataTable =\n","            await google.colab.kernel.invokeFunction('convertToInteractive',\n","                                                     [key], {});\n","          if (!dataTable) return;\n","\n","          const docLinkHtml = 'Like what you see? Visit the ' +\n","            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n","            + ' to learn more about interactive tables.';\n","          element.innerHTML = '';\n","          dataTable['output_type'] = 'display_data';\n","          await google.colab.output.renderOutput(dataTable, element);\n","          const docLink = document.createElement('div');\n","          docLink.innerHTML = docLinkHtml;\n","          element.appendChild(docLink);\n","        }\n","      </script>\n","    </div>\n","  </div>\n","  "],"text/plain":["    iterations  avg_time  time_std           shape       size\n","0          200  0.000459  0.000031            (1,)          1\n","1          200  0.000473  0.000065           (10,)         10\n","2          200  0.000454  0.000026        (10, 10)        100\n","3          200  0.000457  0.000025       (100, 10)       1000\n","4          200  0.000471  0.000056      (100, 100)      10000\n","5          200  0.000461  0.000045     (1000, 100)     100000\n","6          200  0.000458  0.000045    (1000, 1000)    1000000\n","7          200  0.000454  0.000026   (10000, 1000)   10000000\n","8          200  0.000460  0.000058   (10000, 2000)   20000000\n","9          200  0.000464  0.000038   (10000, 5000)   50000000\n","10         200  0.000468  0.000046  (10000, 10000)  100000000"]},"metadata":{},"output_type":"display_data"}],"source":["display(py_data, tf_data, tf_data2)"]},{"cell_type":"code","execution_count":null,"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"executionInfo":{"elapsed":152,"status":"ok","timestamp":1679340682996,"user":{"displayName":"Matt Schoenbauer","userId":"02579049095237637221"},"user_tz":240},"id":"jtqVG8LHT_xe","outputId":"1b5cab19-0c30-46d3-ec16-820566e7000a"},"outputs":[{"data":{"text/plain":["0.0004875003207813609"]},"execution_count":18,"metadata":{},"output_type":"execute_result"}],"source":["tf_data['avg_time'].mean()"]},{"cell_type":"code","execution_count":null,"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"executionInfo":{"elapsed":159,"status":"ok","timestamp":1679340692054,"user":{"displayName":"Matt Schoenbauer","userId":"02579049095237637221"},"user_tz":240},"id":"EEElXDn4gT0T","outputId":"db4d5ccb-ccb2-4cec-d237-e1ea6020d8a7"},"outputs":[{"data":{"text/plain":["0.0011420809138904919"]},"execution_count":19,"metadata":{},"output_type":"execute_result"}],"source":["py_data['avg_time'].mean()"]},{"cell_type":"code","execution_count":null,"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":285},"executionInfo":{"elapsed":611,"status":"ok","timestamp":1679340954310,"user":{"displayName":"Matt Schoenbauer","userId":"02579049095237637221"},"user_tz":240},"id":"OQYlUbjSgV4j","outputId":"6f174d4f-1fb9-4ccd-9bda-a44b72f203b1"},"outputs":[{"data":{"image/png":"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\n","text/plain":["<Figure size 432x288 with 1 Axes>"]},"metadata":{},"output_type":"display_data"}],"source":["plt.clf()\n","plt.plot(tf_data['avg_time'], label='tf')\n","plt.plot(py_data['avg_time'], label='py')\n","plt.plot(tf_data2['avg_time'], label='tf2')\n","plt.legend()\n","plt.show()"]},{"cell_type":"code","execution_count":null,"metadata":{"id":"La0dA4WlXV1T"},"outputs":[],"source":["import tensorflow as tf\n","import numpy as np\n","tf.random.set_seed(0)\n"]},{"cell_type":"code","execution_count":null,"metadata":{"id":"-5MDwCf3ZiH8"},"outputs":[],"source":["x = tf.sort(tf.random.normal((10, )))\n"]},{"cell_type":"code","execution_count":null,"metadata":{"id":"Bp-Gspn4gkez"},"outputs":[],"source":["\n","bits = 8\n","symmetric = True\n","axis = [0]\n","\n","levels = (2 ** (bits - 1) - 1) * 2 if symmetric else (2 ** bits) - 1\n","\n","scale = (tf.keras.backend.max(abs(x), axis=axis, keepdims=True) * 2) / levels\n","\n","\n","def _quantize_x(levels, scale, x):\n","    \"\"\"Get x quantized to the nearest integer, capped at levels/2\"\"\"\n","    v = tf.floor(tf.abs(x) / scale + 0.5)\n","    mask = v < levels / 2\n","    z = tf.sign(x) * tf.where(mask, v, tf.ones_like(v) * levels / 2)\n","    return z\n","\n","def _quantize_x2(levels, scale, x):\n","    \"\"\"Get x quantized to the nearest integer, capped at levels/2\"\"\"\n","    scaled_x = x / scale\n","    clipped_scaled_x = tf.keras.backend.clip(scaled_x, -levels/2, levels/2)\n","    z = tf.round(clipped_scaled_x)\n","    return z\n","\n","\n"]},{"cell_type":"code","execution_count":null,"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"executionInfo":{"elapsed":6,"status":"ok","timestamp":1679406048176,"user":{"displayName":"Matt Schoenbauer","userId":"02579049095237637221"},"user_tz":240},"id":"5z9YC9hZXJEH","outputId":"c8f2ab8b-4fc9-492c-a5c2-d3ea9eaa928f"},"outputs":[{"data":{"text/plain":["array([-127.,  -57.,  -56.,  -32.,  -11.,    6.,   12.,   49.,   65.,\n","        125.], dtype=float32)"]},"execution_count":41,"metadata":{},"output_type":"execute_result"}],"source":["_quantize_x(levels, scale, x).numpy()\n"]},{"cell_type":"code","execution_count":null,"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"executionInfo":{"elapsed":4,"status":"ok","timestamp":1679406048639,"user":{"displayName":"Matt Schoenbauer","userId":"02579049095237637221"},"user_tz":240},"id":"amtJyTsXZR0f","outputId":"d8847c8b-2628-4f40-8f5a-608d5fd96a34"},"outputs":[{"data":{"text/plain":["True"]},"execution_count":42,"metadata":{},"output_type":"execute_result"}],"source":["np.all(_quantize_x2(levels, scale, x).numpy() == _quantize_x(levels, scale, x).numpy())\n"]},{"cell_type":"code","execution_count":null,"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"executionInfo":{"elapsed":393,"status":"ok","timestamp":1679405497809,"user":{"displayName":"Matt Schoenbauer","userId":"02579049095237637221"},"user_tz":240},"id":"KNRasF2AXK_G","outputId":"0e2d6a97-0697-47a5-a1a5-a21853262255"},"outputs":[{"data":{"text/plain":["array([-1.2368279 , -1.0360373 , -0.41969493, -0.01397488,  0.42292207,\n","        0.47027302,  0.5997111 ,  0.60253334,  1.1888583 ,  1.5110626 ],\n","      dtype=float32)"]},"execution_count":31,"metadata":{},"output_type":"execute_result"}],"source":["x.numpy()"]},{"cell_type":"code","execution_count":null,"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"executionInfo":{"elapsed":6,"status":"ok","timestamp":1679405513334,"user":{"displayName":"Matt Schoenbauer","userId":"02579049095237637221"},"user_tz":240},"id":"Lw2Z5Tq_XLQE","outputId":"e6496176-f375-4374-94c1-3b54967fac39"},"outputs":[{"data":{"text/plain":["<tf.Tensor: shape=(1,), dtype=float32, numpy=array([0.01189813], dtype=float32)>"]},"execution_count":32,"metadata":{},"output_type":"execute_result"}],"source":["scale"]},{"cell_type":"code","execution_count":null,"metadata":{"id":"692NdKdZXm-l"},"outputs":[],"source":[]}],"metadata":{"accelerator":"GPU","colab":{"provenance":[],"authorship_tag":"ABX9TyObehyhavgfAD3cp5hmwjOa"},"gpuClass":"standard","kernelspec":{"display_name":"Python 3","name":"python3"},"language_info":{"name":"python"}},"nbformat":4,"nbformat_minor":0}