Day014_BathNormalization_HW.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "from keras.models import Sequential  #用來啟動 NN\n",
    "from keras.layers import Conv2D  # Convolution Operation\n",
    "from keras.layers import MaxPooling2D # Pooling\n",
    "from keras.layers import Flatten\n",
    "from keras.layers import Dense # Fully Connected Networks\n",
    "from keras.layers import BatchNormalization\n",
    "from keras.layers import Activation\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "conv2d_2 (Conv2D)            (None, 32, 32, 32)        896       \n",
      "_________________________________________________________________\n",
      "batch_normalization_2 (Batch (None, 32, 32, 32)        128       \n",
      "_________________________________________________________________\n",
      "activation_2 (Activation)    (None, 32, 32, 32)        0         \n",
      "_________________________________________________________________\n",
      "conv2d_3 (Conv2D)            (None, 32, 32, 32)        9248      \n",
      "_________________________________________________________________\n",
      "batch_normalization_3 (Batch (None, 32, 32, 32)        128       \n",
      "_________________________________________________________________\n",
      "activation_3 (Activation)    (None, 32, 32, 32)        0         \n",
      "=================================================================\n",
      "Total params: 10,400\n",
      "Trainable params: 10,272\n",
      "Non-trainable params: 128\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "input_shape = (32, 32, 3)\n",
    "\n",
    "\n",
    "##  Conv2D-BN-Activation('sigmoid') \n",
    "\n",
    "#def of BN\n",
    "#Normalize the activations of the previous layer at each batch,\n",
    "#i.e. applies a transformation that maintains the mean activation close to 0 \n",
    "#and the activation standard deviation close to 1.\n",
    "\n",
    "#BatchNormalization主要參數：\n",
    "#momentum: Momentum for the moving mean and the moving variance.\n",
    "#epsilon: Small float added to variance to avoid dividing by zero.\n",
    "\n",
    "model = Sequential()\n",
    "\n",
    "##  Activation('sigmoid') 、Activation('relu')\n",
    "\n",
    "model.add(Conv2D(32, kernel_size=(3, 3), padding='same', input_shape=input_shape))\n",
    "model.add(BatchNormalization(momentum=0.99, epsilon=0.001)) \n",
    "model.add(Activation('sigmoid'))\n",
    "\n",
    "model.add(Conv2D(32, kernel_size=(3, 3), padding='same', input_shape=input_shape))\n",
    "model.add(BatchNormalization(momentum=0.99, epsilon=0.001)) \n",
    "model.add(Activation('relu'))\n",
    "\n",
    "\n",
    "\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "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": "3.7.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}