Chainer_CIFAR.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "deletable": true,
    "editable": true
   },
   "source": [
    "# High-level Chainer Example"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": true,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "import os\n",
    "os.environ['CHAINER_TYPE_CHECK'] = '0'\n",
    "\n",
    "import sys\n",
    "import numpy as np\n",
    "import math\n",
    "import chainer\n",
    "import chainer.functions as F\n",
    "import chainer.links as L\n",
    "from chainer import optimizers\n",
    "from chainer import cuda\n",
    "from common.params import *\n",
    "from common.utils import *"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "cuda.set_max_workspace_size(512 * 1024 * 1024)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "OS:  linux\n",
      "Python:  3.5.2 |Anaconda custom (64-bit)| (default, Jul  2 2016, 17:53:06) \n",
      "[GCC 4.4.7 20120313 (Red Hat 4.4.7-1)]\n",
      "Chainer:  2.0.2\n",
      "CuPy:  1.0.2\n",
      "Numpy:  1.13.1\n",
      "GPU:  ['Tesla K80']\n"
     ]
    }
   ],
   "source": [
    "print(\"OS: \", sys.platform)\n",
    "print(\"Python: \", sys.version)\n",
    "print(\"Chainer: \", chainer.__version__)\n",
    "print(\"CuPy: \", chainer.cuda.cupy.__version__)\n",
    "print(\"Numpy: \", np.__version__)\n",
    "print(\"GPU: \", get_gpu_name())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": true,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "class SymbolModule(chainer.Chain):\n",
    "    def __init__(self):\n",
    "        super(SymbolModule, self).__init__()\n",
    "        with self.init_scope():\n",
    "            self.conv1 = L.Convolution2D(3, 50, ksize=3, pad=1)\n",
    "            self.conv2 = L.Convolution2D(50, 50, ksize=3, pad=1)\n",
    "            self.conv3 = L.Convolution2D(50, 100, ksize=3, pad=1)\n",
    "            self.conv4 = L.Convolution2D(100, 100, ksize=3, pad=1)\n",
    "            # feature map size is 8*8 by pooling\n",
    "            self.fc1 = L.Linear(100*8*8, 512)\n",
    "            self.fc2 = L.Linear(512, N_CLASSES)\n",
    "    \n",
    "    def __call__(self, x):\n",
    "        h = F.relu(self.conv2(F.relu(self.conv1(x))))\n",
    "        h = F.max_pooling_2d(h, ksize=2, stride=2)\n",
    "        h = F.dropout(h, 0.25)\n",
    "        \n",
    "        h = F.relu(self.conv4(F.relu(self.conv3(h))))\n",
    "        h = F.max_pooling_2d(h, ksize=2, stride=2)\n",
    "        h = F.dropout(h, 0.25)       \n",
    "        \n",
    "        h = F.dropout(F.relu(self.fc1(h)), 0.5)\n",
    "        return self.fc2(h)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": true,
    "deletable": true,
    "editable": true
   },
   "outputs": [],
   "source": [
    "def init_model(m):\n",
    "    optimizer = optimizers.MomentumSGD(lr=LR, momentum=MOMENTUM)\n",
    "    optimizer.setup(m)\n",
    "    return optimizer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true,
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Downloading http://www.cs.toronto.edu/~kriz/cifar-10-python.tar.gz\n",
      "Done.\n",
      "Extracting files...\n",
      "Done.\n",
      "Preparing train set...\n",
      "Preparing test set...\n",
      "Done.\n",
      "(50000, 3, 32, 32) (10000, 3, 32, 32) (50000,) (10000,)\n",
      "float32 float32 int32 int32\n",
      "CPU times: user 2.78 s, sys: 1.47 s, total: 4.26 s\n",
      "Wall time: 24.4 s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "# Data into format for library\n",
    "#x_train, x_test, y_train, y_test = mnist_for_library(channel_first=True)\n",
    "x_train, x_test, y_train, y_test = cifar_for_library(channel_first=True)\n",
    "print(x_train.shape, x_test.shape, y_train.shape, y_test.shape)\n",
    "print(x_train.dtype, x_test.dtype, y_train.dtype, y_test.dtype)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 188 ms, sys: 324 ms, total: 512 ms\n",
      "Wall time: 544 ms\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "# Create symbol\n",
    "sym = SymbolModule()\n",
    "if GPU:\n",
    "    chainer.cuda.get_device(0).use()  # Make a specified GPU current\n",
    "    sym.to_gpu()  # Copy the model to the GPU"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 0 ns, sys: 0 ns, total: 0 ns\n",
      "Wall time: 89.4 µs\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "optimizer = init_model(sym)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0\n",
      "1\n",
      "2\n",
      "3\n",
      "4\n",
      "5\n",
      "6\n",
      "7\n",
      "8\n",
      "9\n",
      "CPU times: user 3min 25s, sys: 34.9 s, total: 4min\n",
      "Wall time: 4min\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "for j in range(EPOCHS):\n",
    "    for data, target in yield_mb(x_train, y_train, BATCHSIZE, shuffle=True):\n",
    "        # Get samples\n",
    "        data = cuda.to_gpu(data)\n",
    "        target = cuda.to_gpu(target)\n",
    "        output = sym(data)\n",
    "        loss = F.softmax_cross_entropy(output, target)\n",
    "        sym.cleargrads()\n",
    "        loss.backward()\n",
    "        optimizer.update()\n",
    "    # Log\n",
    "    print(j)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 1.32 s, sys: 308 ms, total: 1.63 s\n",
      "Wall time: 1.63 s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "n_samples = (y_test.shape[0]//BATCHSIZE)*BATCHSIZE\n",
    "y_guess = np.zeros(n_samples, dtype=np.int)\n",
    "y_truth = y_test[:n_samples]\n",
    "c = 0\n",
    "\n",
    "with chainer.using_config('train', False), chainer.using_config('enable_backprop', False):\n",
    "    for data, target in yield_mb(x_test, y_test, BATCHSIZE):\n",
    "        # Forwards\n",
    "        pred = cuda.to_cpu(sym(cuda.to_gpu(data)).data.argmax(-1))\n",
    "        # Collect results\n",
    "        y_guess[c*BATCHSIZE:(c+1)*BATCHSIZE] = pred\n",
    "        c += 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": false,
    "deletable": true,
    "editable": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy:  0.788762019231\n"
     ]
    }
   ],
   "source": [
    "print(\"Accuracy: \", sum(y_guess == y_truth)/len(y_guess))"
   ]
  }
 ],
 "metadata": {
  "anaconda-cloud": {},
  "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": "3.5.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}