AlexNet TensorFlow 2.0.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "%reload_ext autoreload\n",
    "%autoreload 2\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "import tensorflow as tf\n",
    "import tensorflow.keras as keras\n",
    "from tensorflow.keras import layers\n",
    "# from tensorflow.keras.datasets import cifar10\n",
    "# import numpy as np"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "class AlexBlockA(keras.Model):\n",
    "    def __init__(self, nf, kernel=(3, 3), cs=1, ps=2, input_shape=None, activation='relu', data_format='channels_first'):\n",
    "        super(AlexBlockA, self).__init__(name='AlexBlockA')\n",
    "        self.padding = layers.ZeroPadding2D(2, data_format=data_format)\n",
    "        if input_shape and input_shape is not None: self.conv = layers.Conv2D(nf, kernel, strides=(cs, cs), input_shape=input_shape, activation=activation, data_format=data_format)\n",
    "        else: self.conv = layers.Conv2D(nf, kernel, strides=(cs, cs), activation=activation, data_format=data_format)\n",
    "        self.pool = layers.MaxPool2D(3, strides=ps, data_format=data_format)\n",
    "    def call(self, x):\n",
    "        x = self.padding(x)\n",
    "        x = self.conv(x)\n",
    "        x = self.pool(x)\n",
    "        print('from alexa', x.shape)\n",
    "        return x"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "class AlexBlockB(keras.Model):\n",
    "    \"\"\"\n",
    "    nfs: filter for each conv layer\n",
    "    kernels: kernel for each conv layer\n",
    "    sts: stride for each conv layer\n",
    "    \"\"\"\n",
    "    def __init__(self, nfs, kernels, sts, pool_size=(3, 3), pool_stride=1, activation='relu', padding='same'):\n",
    "        super(AlexBlockB, self).__init__(name='AlexBlockB')\n",
    "        if isinstance(nfs, int): nfs = [nfs for _ in range(3)]\n",
    "        if isinstance(sts, int): sts = [sts for _ in range(3)]\n",
    "        if len(kernels) == 1: kernels = [kernels[0] for _ in range(3)]\n",
    "        self.conv = []\n",
    "        for i in range(3):\n",
    "            if i == 2: self.conv.append(layers.ZeroPadding2D(2))\n",
    "            self.conv.append(layers.Conv2D(nfs[i], kernels[i], strides=sts[i], activation=activation))\n",
    "        self.pool = layers.MaxPool2D(pool_size, strides=pool_stride, padding=padding)\n",
    "    def call(self, x):\n",
    "        for i in range(4):\n",
    "            x = self.conv[i](x)\n",
    "        x = self.pool(x)\n",
    "        print('from alexb', x.shape)\n",
    "        return x"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "class AlexNet(keras.Model):\n",
    "    def __init__(self, input_shape, data_format='channels_first'):\n",
    "        super(AlexNet, self).__init__(name='AlexNet')\n",
    "        tf.keras.backend.set_image_data_format(data_format)\n",
    "        self.body = keras.Sequential(name='Body')\n",
    "        self.body.add(AlexBlockA(96, (11, 11), 4, 2, input_shape))\n",
    "        self.body.add(AlexBlockA(256, (5, 5), 1, 2))\n",
    "        self.body.add(AlexBlockB([384, 384, 256], [(3, 3)], [1, 1, 2]))\n",
    "        self.fc1 = layers.Dense(4096, activation='relu')\n",
    "        self.fc2 = layers.Dense(4096, activation='relu')\n",
    "        self.fc3 = layers.Dense(1000, activation='softmax')\n",
    "    def call(self, x):\n",
    "        print(x.shape)\n",
    "        x = self.body(x)\n",
    "        x = layers.Flatten()(x)\n",
    "        print('from flatten', x.shape)\n",
    "        x = self.fc1(x)\n",
    "        print('from dense 4096', x.shape)\n",
    "        x = self.fc2(x)\n",
    "        print('from dense 4096', x.shape)\n",
    "        x = self.fc3(x)\n",
    "        print('from dense 1000', x.shape)\n",
    "        return x"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "ename": "NameError",
     "evalue": "name 'data_format' is not defined",
     "output_type": "error",
     "traceback": [
      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[1;31mNameError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[1;32m<ipython-input-6-7dcdc64c4d32>\u001b[0m in \u001b[0;36m<module>\u001b[1;34m\u001b[0m\n\u001b[1;32m----> 1\u001b[1;33m \u001b[0malex\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mAlexNet\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0minput_shape\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;36m3\u001b[0m\u001b[1;33m,\u001b[0m \u001b[1;36m224\u001b[0m\u001b[1;33m,\u001b[0m \u001b[1;36m224\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m      2\u001b[0m \u001b[0malex\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mbuild\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;36m1\u001b[0m\u001b[1;33m,\u001b[0m \u001b[1;36m3\u001b[0m\u001b[1;33m,\u001b[0m \u001b[1;36m224\u001b[0m\u001b[1;33m,\u001b[0m \u001b[1;36m224\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m      3\u001b[0m \u001b[0malex\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0msummary\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32m<ipython-input-5-6eb9f27d5fd4>\u001b[0m in \u001b[0;36m__init__\u001b[1;34m(self, input_shape, data_format)\u001b[0m\n\u001b[0;32m      6\u001b[0m         \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mbody\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0madd\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mAlexBlockA\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;36m96\u001b[0m\u001b[1;33m,\u001b[0m \u001b[1;33m(\u001b[0m\u001b[1;36m11\u001b[0m\u001b[1;33m,\u001b[0m \u001b[1;36m11\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m,\u001b[0m \u001b[1;36m4\u001b[0m\u001b[1;33m,\u001b[0m \u001b[1;36m2\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0minput_shape\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m      7\u001b[0m         \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mbody\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0madd\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mAlexBlockA\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;36m256\u001b[0m\u001b[1;33m,\u001b[0m \u001b[1;33m(\u001b[0m\u001b[1;36m5\u001b[0m\u001b[1;33m,\u001b[0m \u001b[1;36m5\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m,\u001b[0m \u001b[1;36m1\u001b[0m\u001b[1;33m,\u001b[0m \u001b[1;36m2\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m----> 8\u001b[1;33m         \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mbody\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0madd\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mAlexBlockB\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m[\u001b[0m\u001b[1;36m384\u001b[0m\u001b[1;33m,\u001b[0m \u001b[1;36m384\u001b[0m\u001b[1;33m,\u001b[0m \u001b[1;36m256\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m,\u001b[0m \u001b[1;33m[\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;36m3\u001b[0m\u001b[1;33m,\u001b[0m \u001b[1;36m3\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m,\u001b[0m \u001b[1;33m[\u001b[0m\u001b[1;36m1\u001b[0m\u001b[1;33m,\u001b[0m \u001b[1;36m1\u001b[0m\u001b[1;33m,\u001b[0m \u001b[1;36m2\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m      9\u001b[0m         \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mfc1\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mlayers\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mDense\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;36m4096\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mactivation\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;34m'relu'\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m     10\u001b[0m         \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mfc2\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mlayers\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mDense\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;36m4096\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mactivation\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;34m'relu'\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32m<ipython-input-4-2d531d50b5b9>\u001b[0m in \u001b[0;36m__init__\u001b[1;34m(self, nfs, kernels, sts, pool_size, pool_stride, activation, padding)\u001b[0m\n\u001b[0;32m     13\u001b[0m         \u001b[1;32mfor\u001b[0m \u001b[0mi\u001b[0m \u001b[1;32min\u001b[0m \u001b[0mrange\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;36m3\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m     14\u001b[0m             \u001b[1;32mif\u001b[0m \u001b[0mi\u001b[0m \u001b[1;33m==\u001b[0m \u001b[1;36m2\u001b[0m\u001b[1;33m:\u001b[0m \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mconv\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mappend\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mlayers\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mZeroPadding2D\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;36m2\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m---> 15\u001b[1;33m             \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mconv\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mappend\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mlayers\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mConv2D\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mnfs\u001b[0m\u001b[1;33m[\u001b[0m\u001b[0mi\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mkernels\u001b[0m\u001b[1;33m[\u001b[0m\u001b[0mi\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mstrides\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0msts\u001b[0m\u001b[1;33m[\u001b[0m\u001b[0mi\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mactivation\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mactivation\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mdata_format\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mdata_format\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m     16\u001b[0m         \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mpool\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mlayers\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mMaxPool2D\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mpool_size\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mstrides\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mpool_stride\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mpadding\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mpadding\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m     17\u001b[0m     \u001b[1;32mdef\u001b[0m \u001b[0mcall\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mx\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;31mNameError\u001b[0m: name 'data_format' is not defined"
     ]
    }
   ],
   "source": [
    "alex = AlexNet(input_shape=(3, 224, 224))\n",
    "alex.build((1, 3, 224, 224))\n",
    "alex.summary()"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}