deep_learning_mnist.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "cellView": "both",
    "colab": {
     "autoexec": {
      "startup": false,
      "wait_interval": 0
     }
    },
    "colab_type": "code",
    "id": "apJbCsBHl-2A"
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "#Choose GPUs and CPUs\n",
    "\n",
    "#If want to choose GPU to use\n",
    "import os\n",
    "os.environ[\"CUDA_VISIBLE_DEVICES\"]=\"1,4\"\n",
    "#This will make it so that only the 0th GPU is visible to TensorFlow.\n",
    "#In your case, you can choose any in the range [0, 3].\n",
    "#If you wanted, for example, the 0th and 2nd GPUs to both be visible, replace \"0\" with \"0,2\"\n",
    "\n",
    "import tensorflow as tf\n",
    "import keras\n",
    "from keras import backend as K\n",
    "from keras.datasets import mnist\n",
    "\n",
    "\n",
    "GPU=True\n",
    "CPU=False\n",
    "num_cores = 4\n",
    "\n",
    "if GPU:\n",
    "    num_GPU = 2\n",
    "    num_CPU = 1\n",
    "if CPU:\n",
    "    num_CPU = 1\n",
    "    num_GPU = 0\n",
    "\n",
    "config = tf.ConfigProto(intra_op_parallelism_threads=num_cores,\\\n",
    "        inter_op_parallelism_threads=num_cores, allow_soft_placement=True,\\\n",
    "        device_count = {'CPU' : num_CPU, 'GPU' : num_GPU},\\\n",
    "#        gpu_options=tf.GPUOptions(per_process_gpu_memory_fraction=0.5))\n",
    "                       )\n",
    "session = tf.Session(config=config)\n",
    "K.set_session(session)\n",
    "# These are all the modules we'll be using later. Make sure you can import them\n",
    "# before proceeding further.\n",
    "from __future__ import print_function\n",
    "import imageio\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import os\n",
    "import sys\n",
    "import tarfile\n",
    "from IPython.display import display, Image\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from six.moves.urllib.request import urlretrieve\n",
    "from six.moves import cPickle as pickle\n",
    "\n",
    "# Config the matplotlib backend as plotting inline in IPython\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from keras import Model\n",
    "from keras.utils import multi_gpu_model\n",
    "\n",
    "\n",
    "class ModelMGPU(Model):\n",
    "    def __init__(self, ser_model, gpus):\n",
    "        pmodel = multi_gpu_model(ser_model, gpus)\n",
    "        self.__dict__.update(pmodel.__dict__)\n",
    "        self._smodel = ser_model\n",
    "\n",
    "    def __getattribute__(self, attrname):\n",
    "        '''Override load and save methods to be used from the serial-model. The\n",
    "        serial-model holds references to the weights in the multi-gpu model.\n",
    "        '''\n",
    "        # return Model.__getattribute__(self, attrname)\n",
    "        if 'load' in attrname or 'save' in attrname:\n",
    "            return getattr(self._smodel, attrname)\n",
    "\n",
    "        return super(ModelMGPU, self).__getattribute__(attrname)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Load MNIST data from Keras"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train_dataset shape: (60000, 28, 28, 1)\n",
      "60000 train samples\n",
      "10000 test samples\n"
     ]
    }
   ],
   "source": [
    "\n",
    "# input image dimensions\n",
    "img_rows, img_cols = 28, 28\n",
    "num_classes=10\n",
    "\n",
    "# the data, split between train and test sets\n",
    "(x_train, y_train), (x_test, y_test) = mnist.load_data()\n",
    "\n",
    "if K.image_data_format() == 'channels_first':\n",
    "    x_train = x_train.reshape(x_train.shape[0], 1, img_rows, img_cols)\n",
    "    x_test = x_test.reshape(x_test.shape[0], 1, img_rows, img_cols)\n",
    "    input_shape = (1, img_rows, img_cols)\n",
    "else:\n",
    "    x_train = x_train.reshape(x_train.shape[0], img_rows, img_cols, 1)\n",
    "    x_test = x_test.reshape(x_test.shape[0], img_rows, img_cols, 1)\n",
    "    input_shape = (img_rows, img_cols, 1)\n",
    "\n",
    "train_dataset = x_train.astype('float32')\n",
    "test_dataset = x_test.astype('float32')\n",
    "train_dataset /= 255\n",
    "test_dataset /= 255\n",
    "print('train_dataset shape:', x_train.shape)\n",
    "print(x_train.shape[0], 'train samples')\n",
    "print(x_test.shape[0], 'test samples')\n",
    "\n",
    "# convert class vectors to binary class matrices\n",
    "train_labels = keras.utils.to_categorical(y_train, num_classes)\n",
    "test_labels = keras.utils.to_categorical(y_test, num_classes)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# FT the training and test data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(60000, 28, 28, 1)\n",
      "(10000, 28, 28, 1)\n"
     ]
    }
   ],
   "source": [
    "from scipy.fftpack import fftn,fftshift\n",
    "dims=train_dataset.shape\n",
    "ft_train=[]\n",
    "for i in range(dims[0]):\n",
    "    ft_train.append(np.abs(fftshift(fftn(train_dataset[i,:,:]))))\n",
    "\n",
    "dims=test_dataset.shape\n",
    "ft_test=[]\n",
    "for i in range(dims[0]):\n",
    "    ft_test.append(np.abs(fftshift(fftn(test_dataset[i,:,:])))) #Don't forget to take abs!\n",
    "\n",
    "ft_train=np.asarray(ft_train)\n",
    "ft_test=np.asarray(ft_test)\n",
    "print (ft_train.shape)\n",
    "print (ft_test.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x7fd112cdb890>"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAeQAAADtCAYAAABu1gaFAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAKAAJGQAAApAQgYAoAAkZAAACvC/rwAq4VKrV6wAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fd1d8152190>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "f,axarr=plt.subplots(1,2,figsize=(8, 6))\n",
    "axarr[0].imshow(train_dataset[1].squeeze())\n",
    "axarr[1].imshow(ft_train[1].squeeze())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# O.K now try CNN"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(60000, 28, 28, 1)\n",
      "(60000, 28, 28, 1)\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "conv2d_1 (Conv2D)            (None, 26, 26, 32)        320       \n",
      "_________________________________________________________________\n",
      "conv2d_2 (Conv2D)            (None, 24, 24, 32)        9248      \n",
      "_________________________________________________________________\n",
      "max_pooling2d_1 (MaxPooling2 (None, 12, 12, 32)        0         \n",
      "_________________________________________________________________\n",
      "conv2d_3 (Conv2D)            (None, 10, 10, 64)        18496     \n",
      "_________________________________________________________________\n",
      "conv2d_4 (Conv2D)            (None, 8, 8, 64)          36928     \n",
      "_________________________________________________________________\n",
      "max_pooling2d_2 (MaxPooling2 (None, 4, 4, 64)          0         \n",
      "_________________________________________________________________\n",
      "flatten_1 (Flatten)          (None, 1024)              0         \n",
      "_________________________________________________________________\n",
      "dense_1 (Dense)              (None, 1200)              1230000   \n",
      "_________________________________________________________________\n",
      "dropout_1 (Dropout)          (None, 1200)              0         \n",
      "_________________________________________________________________\n",
      "dense_2 (Dense)              (None, 10)                12010     \n",
      "=================================================================\n",
      "Total params: 1,307,002\n",
      "Trainable params: 1,307,002\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n",
      "None\n",
      "__________________________________________________________________________________________________\n",
      "Layer (type)                    Output Shape         Param #     Connected to                     \n",
      "==================================================================================================\n",
      "conv2d_1_input (InputLayer)     (None, 28, 28, 1)    0                                            \n",
      "__________________________________________________________________________________________________\n",
      "lambda_1 (Lambda)               (None, 28, 28, 1)    0           conv2d_1_input[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "lambda_2 (Lambda)               (None, 28, 28, 1)    0           conv2d_1_input[0][0]             \n",
      "__________________________________________________________________________________________________\n",
      "sequential_1 (Sequential)       (None, 10)           1307002     lambda_1[0][0]                   \n",
      "                                                                 lambda_2[0][0]                   \n",
      "__________________________________________________________________________________________________\n",
      "dense_2 (Concatenate)           (None, 10)           0           sequential_1[1][0]               \n",
      "                                                                 sequential_1[2][0]               \n",
      "==================================================================================================\n",
      "Total params: 1,307,002\n",
      "Trainable params: 1,307,002\n",
      "Non-trainable params: 0\n",
      "__________________________________________________________________________________________________\n",
      "None\n",
      "Train on 54000 samples, validate on 6000 samples\n",
      "Epoch 1/10\n",
      "53504/54000 [============================>.] - ETA: 0s - loss: 0.5919 - acc: 0.8095\n",
      "Epoch 00001: saving model to tmp_models/weights.01.hdf5\n",
      "54000/54000 [==============================] - 6s 112us/step - loss: 0.5887 - acc: 0.8105 - val_loss: 0.2211 - val_acc: 0.9260\n",
      "Epoch 2/10\n",
      "53760/54000 [============================>.] - ETA: 0s - loss: 0.2260 - acc: 0.9266\n",
      "Epoch 00002: saving model to tmp_models/weights.02.hdf5\n",
      "54000/54000 [==============================] - 6s 108us/step - loss: 0.2257 - acc: 0.9267 - val_loss: 0.1661 - val_acc: 0.9462\n",
      "Epoch 3/10\n",
      "53760/54000 [============================>.] - ETA: 0s - loss: 0.1755 - acc: 0.9413- ETA: 0s - loss: 0.1770 - acc: 0.\n",
      "Epoch 00003: saving model to tmp_models/weights.03.hdf5\n",
      "54000/54000 [==============================] - 6s 109us/step - loss: 0.1752 - acc: 0.9414 - val_loss: 0.1616 - val_acc: 0.9468\n",
      "Epoch 4/10\n",
      "53760/54000 [============================>.] - ETA: 0s - loss: 0.1498 - acc: 0.9497\n",
      "Epoch 00004: saving model to tmp_models/weights.04.hdf5\n",
      "54000/54000 [==============================] - 6s 107us/step - loss: 0.1498 - acc: 0.9497 - val_loss: 0.1511 - val_acc: 0.9503\n",
      "Epoch 5/10\n",
      "53504/54000 [============================>.] - ETA: 0s - loss: 0.1282 - acc: 0.9568\n",
      "Epoch 00005: saving model to tmp_models/weights.05.hdf5\n",
      "54000/54000 [==============================] - 6s 107us/step - loss: 0.1279 - acc: 0.9569 - val_loss: 0.1365 - val_acc: 0.9533\n",
      "Epoch 6/10\n",
      "53760/54000 [============================>.] - ETA: 0s - loss: 0.1134 - acc: 0.9610\n",
      "Epoch 00006: saving model to tmp_models/weights.06.hdf5\n",
      "54000/54000 [==============================] - 6s 108us/step - loss: 0.1134 - acc: 0.9611 - val_loss: 0.1490 - val_acc: 0.9507\n",
      "Epoch 7/10\n",
      "53504/54000 [============================>.] - ETA: 0s - loss: 0.1000 - acc: 0.9656\n",
      "Epoch 00007: saving model to tmp_models/weights.07.hdf5\n",
      "54000/54000 [==============================] - 6s 107us/step - loss: 0.1002 - acc: 0.9655 - val_loss: 0.1349 - val_acc: 0.9560\n",
      "Epoch 8/10\n",
      "53760/54000 [============================>.] - ETA: 0s - loss: 0.0897 - acc: 0.9682\n",
      "Epoch 00008: saving model to tmp_models/weights.08.hdf5\n",
      "54000/54000 [==============================] - 6s 108us/step - loss: 0.0897 - acc: 0.9682 - val_loss: 0.1344 - val_acc: 0.9592\n",
      "Epoch 9/10\n",
      "53504/54000 [============================>.] - ETA: 0s - loss: 0.0808 - acc: 0.9717\n",
      "Epoch 00009: saving model to tmp_models/weights.09.hdf5\n",
      "54000/54000 [==============================] - 6s 109us/step - loss: 0.0813 - acc: 0.9715 - val_loss: 0.1388 - val_acc: 0.9587\n",
      "Epoch 10/10\n",
      "53760/54000 [============================>.] - ETA: 0s - loss: 0.0755 - acc: 0.9736\n",
      "Epoch 00010: saving model to tmp_models/weights.10.hdf5\n",
      "54000/54000 [==============================] - 6s 108us/step - loss: 0.0754 - acc: 0.9736 - val_loss: 0.1505 - val_acc: 0.9580\n"
     ]
    }
   ],
   "source": [
    "dims=ft_train.shape\n",
    "print (dims)\n",
    "trainee=ft_train.reshape(dims[0],dims[1],dims[2],1)\n",
    "dims=test_dataset.shape\n",
    "testee=ft_test.reshape(dims[0],dims[1],dims[2],1)\n",
    "print (trainee.shape)\n",
    "\n",
    "\n",
    "from keras.layers import Conv2D, MaxPool2D\n",
    "from keras.models import Sequential\n",
    "from keras.layers import Dense, Dropout, Activation, Flatten\n",
    "from keras import backend as K\n",
    "from keras.utils import to_categorical\n",
    "\n",
    "K.clear_session()\n",
    "np.random.seed(123)\n",
    "model=Sequential()\n",
    "\n",
    "model.add(Conv2D(32,(3,3), activation='relu', input_shape=(28,28,1),data_format='channels_last'))\n",
    "model.add(Conv2D(32,(3,3), activation='relu'))\n",
    "model.add(MaxPool2D(pool_size=(2,2)))\n",
    "model.add(Conv2D(64,(3,3), activation='relu'))\n",
    "model.add(Conv2D(64,(3,3), activation='relu'))\n",
    "model.add(MaxPool2D(pool_size=(2,2)))\n",
    "model.add(Flatten())\n",
    "model.add(Dense(1200, activation='relu'))\n",
    "model.add(Dropout(0.5))\n",
    "model.add(Dense(10, activation='softmax'))\n",
    "\n",
    "parallel_model = ModelMGPU(model, gpus=num_GPU)\n",
    "parallel_model.compile(loss='categorical_crossentropy', optimizer='adam', metrics=['accuracy'])\n",
    "\n",
    "print (model.summary())\n",
    "print (parallel_model.summary())\n",
    "\n",
    "\n",
    "\n",
    "checkpoints=keras.callbacks.ModelCheckpoint('tmp_models/weights.{epoch:02d}.hdf5', monitor='val_loss', \n",
    "                                verbose=1, save_best_only=False, save_weights_only=False, mode='auto', period=1)\n",
    "\n",
    "history=parallel_model.fit(trainee, train_labels, batch_size=256, verbose=1, epochs=10, validation_split = 0.1, callbacks=[checkpoints])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAwwAAAGZCAYAAAAkS7aiAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n4mGSXlMoJ1wAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fd0e5a58e10>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "hist=history\n",
    "epochs=np.asarray(history.epoch)+1\n",
    "\n",
    "%matplotlib inline\n",
    "import matplotlib\n",
    "plt.style.use('seaborn-white')\n",
    "matplotlib.rc('font',family='Times New Roman')\n",
    "matplotlib.rcParams['font.size'] = 20\n",
    "\n",
    "f, axarr = plt.subplots(2, sharex=True, figsize=(8, 6))\n",
    "\n",
    "axarr[0].set(ylabel='Loss')\n",
    "axarr[0].plot(epochs,hist.history['loss'], 'C3o', label='Training')\n",
    "axarr[0].plot(epochs,hist.history['val_loss'], 'C3-', label='Validation')\n",
    "axarr[0].grid()\n",
    "axarr[0].legend(loc='center right', bbox_to_anchor=(1.5, 0.5))\n",
    "\n",
    "\n",
    "axarr[1].plot(epochs,hist.history['acc'], 'C0o', label='Training')\n",
    "axarr[1].plot(epochs,hist.history['val_acc'], 'C0-', label='Validation')\n",
    "plt.legend(loc='center right', bbox_to_anchor=(1.5, 0.5))\n",
    "plt.xlabel('Epochs')\n",
    "plt.ylabel('Accuracy')\n",
    "plt.tight_layout()\n",
    "plt.grid()\n",
    "plt.tick_params(which='both', width=2)\n",
    "plt.tick_params(which='major', length=7)\n",
    "plt.tick_params(which='minor', length=5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "10000/10000 [==============================] - 1s 143us/step\n",
      "[0.14821517372876405, 0.95540000000000003]\n"
     ]
    }
   ],
   "source": [
    "from keras.models import Model, load_model\n",
    "model=load_model('tmp_models/weights.08.hdf5')\n",
    "model.compile(loss='categorical_crossentropy', optimizer='adam', metrics=['accuracy'])\n",
    "score=model.evaluate(testee, test_labels, verbose=1)\n",
    "print (score)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "preds=(model.predict(testee))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "6\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAeQAAADyCAYAAACcVrbLAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fd0e7720410>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "i=100\n",
    "f,axarr=plt.subplots(1,2,figsize=(8, 6))\n",
    "axarr[0].imshow(test_dataset[i].squeeze(),cmap='viridis')\n",
    "axarr[1].imshow(testee[i].squeeze(), cmap='viridis')\n",
    "print (np.argmax(preds[i]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "colab": {
   "default_view": {},
   "name": "1_notmnist.ipynb",
   "provenance": [],
   "version": "0.3.2",
   "views": {}
  },
  "kernelspec": {
   "display_name": "Python 2",
   "language": "python",
   "name": "python2"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.14"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}