number_parameters_dense_conv_nn.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "https://medium.com/p/3eb35ab7f3c/edit"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import utilities"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "img_rows, img_cols = 224, 224\n",
    "colors = 3\n",
    "input_size = img_rows * img_cols * colors\n",
    "input_shape = (img_rows, img_cols, colors)\n",
    "\n",
    "num_classes = 10"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "dense_1 (Dense)              (None, 32)                4816928   \n",
      "_________________________________________________________________\n",
      "dense_2 (Dense)              (None, 64)                2112      \n",
      "_________________________________________________________________\n",
      "dense_3 (Dense)              (None, 128)               8320      \n",
      "_________________________________________________________________\n",
      "dense_4 (Dense)              (None, 10)                1290      \n",
      "=================================================================\n",
      "Total params: 4,828,650\n",
      "Trainable params: 4,828,650\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "from keras.models import Sequential\n",
    "from keras.layers import Dense\n",
    "\n",
    "model = Sequential([\n",
    "    Dense(32, activation='relu', input_shape=(input_size,)),\n",
    "    Dense(64, activation='relu'),\n",
    "    Dense(128, activation='relu'),\n",
    "    Dense(num_classes, activation='softmax')\n",
    "])\n",
    "\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "-------------\n",
      "(150528, 32)\n",
      "(32,)\n",
      "-------------\n",
      "(32, 64)\n",
      "(64,)\n",
      "-------------\n",
      "(64, 128)\n",
      "(128,)\n",
      "-------------\n",
      "(128, 10)\n",
      "(10,)\n"
     ]
    }
   ],
   "source": [
    "utilities.print_weights_shape(model)"
   ]
  },
  {
   "cell_type": "raw",
   "metadata": {},
   "source": [
    "output_size * (input_size + 1) == number_parameters"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "assert 32 * (input_size + 1) == 4816928\n",
    "assert 64 * (32 + 1) == 2112\n",
    "assert 128 * (64 + 1) == 8320\n",
    "assert num_classes * (128 + 1) == 1290"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "conv2d_1 (Conv2D)            (None, 222, 222, 32)      896       \n",
      "_________________________________________________________________\n",
      "conv2d_2 (Conv2D)            (None, 220, 220, 64)      18496     \n",
      "_________________________________________________________________\n",
      "conv2d_3 (Conv2D)            (None, 218, 218, 128)     73856     \n",
      "_________________________________________________________________\n",
      "dense_5 (Dense)              (None, 218, 218, 10)      1290      \n",
      "=================================================================\n",
      "Total params: 94,538\n",
      "Trainable params: 94,538\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "from keras.models import Sequential\n",
    "from keras.layers import Conv2D\n",
    "\n",
    "\n",
    "model = Sequential([\n",
    "    Conv2D(32, (3, 3), activation='relu', input_shape=input_shape),\n",
    "    Conv2D(64, (3, 3), activation='relu'),\n",
    "    Conv2D(128, (3, 3), activation='relu'),\n",
    "    Dense(num_classes, activation='softmax')\n",
    "])\n",
    "\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "-------------\n",
      "(3, 3, 3, 32)\n",
      "(32,)\n",
      "-------------\n",
      "(3, 3, 32, 64)\n",
      "(64,)\n",
      "-------------\n",
      "(3, 3, 64, 128)\n",
      "(128,)\n",
      "-------------\n",
      "(128, 10)\n",
      "(10,)\n"
     ]
    }
   ],
   "source": [
    "utilities.print_weights_shape(model)"
   ]
  },
  {
   "cell_type": "raw",
   "metadata": {},
   "source": [
    "output_size * (input_size + 1) == number_parameters\n",
    "\n",
    "i.e.,\n",
    "\n",
    "output_channels * (input_channels * window_size + 1) == number_parameters\n",
    "\n",
    "here window_size=3*3"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "assert 32 * (3 * (3*3) + 1) == 896\n",
    "assert 64 * (32 * (3*3) + 1) == 18496\n",
    "assert 128 * (64 * (3*3) + 1) == 73856\n",
    "assert num_classes * (128 + 1) == 1290"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "conv2d_4 (Conv2D)            (None, 222, 222, 32)      896       \n",
      "_________________________________________________________________\n",
      "conv2d_5 (Conv2D)            (None, 220, 220, 64)      18496     \n",
      "_________________________________________________________________\n",
      "max_pooling2d_1 (MaxPooling2 (None, 110, 110, 64)      0         \n",
      "_________________________________________________________________\n",
      "dropout_1 (Dropout)          (None, 110, 110, 64)      0         \n",
      "_________________________________________________________________\n",
      "flatten_1 (Flatten)          (None, 774400)            0         \n",
      "_________________________________________________________________\n",
      "dense_6 (Dense)              (None, 128)               99123328  \n",
      "_________________________________________________________________\n",
      "dropout_2 (Dropout)          (None, 128)               0         \n",
      "_________________________________________________________________\n",
      "dense_7 (Dense)              (None, 10)                1290      \n",
      "=================================================================\n",
      "Total params: 99,144,010\n",
      "Trainable params: 99,144,010\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "from keras.datasets import mnist\n",
    "from keras.models import Sequential\n",
    "from keras.layers import Dense, Dropout, Flatten\n",
    "from keras.layers import Conv2D, MaxPooling2D\n",
    "\n",
    "model = Sequential()\n",
    "model.add(Conv2D(32, kernel_size=(3, 3),\n",
    "                 activation='relu',\n",
    "                 input_shape=input_shape))\n",
    "model.add(Conv2D(64, (3, 3), activation='relu'))\n",
    "model.add(MaxPooling2D(pool_size=(2, 2)))\n",
    "model.add(Dropout(0.25))\n",
    "model.add(Flatten())\n",
    "model.add(Dense(128, activation='relu'))\n",
    "model.add(Dropout(0.5))\n",
    "model.add(Dense(num_classes, activation='softmax'))\n",
    "\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "-------------\n",
      "(3, 3, 3, 32)\n",
      "(32,)\n",
      "-------------\n",
      "(3, 3, 32, 64)\n",
      "(64,)\n",
      "-------------\n",
      "-------------\n",
      "-------------\n",
      "-------------\n",
      "(774400, 128)\n",
      "(128,)\n",
      "-------------\n",
      "-------------\n",
      "(128, 10)\n",
      "(10,)\n"
     ]
    }
   ],
   "source": [
    "utilities.print_weights_shape(model)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "assert 32 * (3 * (3*3) + 1) == 896\n",
    "assert 64 * (32 * (3*3) + 1) == 18496\n",
    "assert 110 * 110 * 64 == 774400\n",
    "assert 128 * (774400 + 1) == 99123328\n",
    "assert num_classes * (128 + 1) == 1290"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "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": "3.6.1"
  },
  "toc": {
   "nav_menu": {
    "height": "101px",
    "width": "251px"
   },
   "number_sections": true,
   "sideBar": true,
   "skip_h1_title": false,
   "toc_cell": false,
   "toc_position": {
    "height": "394px",
    "left": "0px",
    "right": "866.8px",
    "top": "107px",
    "width": "158px"
   },
   "toc_section_display": "block",
   "toc_window_display": true
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}