softmax.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Softmax exercise\n",
    "\n",
    "*Complete and hand in this completed worksheet (including its outputs and any supporting code outside of the worksheet) with your assignment submission. For more details see the [assignments page](http://vision.stanford.edu/teaching/cs231n/assignments.html) on the course website.*\n",
    "\n",
    "This exercise is analogous to the SVM exercise. You will:\n",
    "\n",
    "- implement a fully-vectorized **loss function** for the Softmax classifier\n",
    "- implement the fully-vectorized expression for its **analytic gradient**\n",
    "- **check your implementation** with numerical gradient\n",
    "- use a validation set to **tune the learning rate and regularization** strength\n",
    "- **optimize** the loss function with **SGD**\n",
    "- **visualize** the final learned weights\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The autoreload extension is already loaded. To reload it, use:\n",
      "  %reload_ext autoreload\n"
     ]
    }
   ],
   "source": [
    "import random\n",
    "import numpy as np\n",
    "from cs231n.data_utils import load_CIFAR10\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "from __future__ import print_function\n",
    "\n",
    "%matplotlib inline\n",
    "plt.rcParams['figure.figsize'] = (10.0, 8.0) # set default size of plots\n",
    "plt.rcParams['image.interpolation'] = 'nearest'\n",
    "plt.rcParams['image.cmap'] = 'gray'\n",
    "\n",
    "# for auto-reloading extenrnal modules\n",
    "# see http://stackoverflow.com/questions/1907993/autoreload-of-modules-in-ipython\n",
    "%load_ext autoreload\n",
    "%autoreload 2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train data shape:  (49000, 3073)\n",
      "Train labels shape:  (49000,)\n",
      "Validation data shape:  (1000, 3073)\n",
      "Validation labels shape:  (1000,)\n",
      "Test data shape:  (1000, 3073)\n",
      "Test labels shape:  (1000,)\n",
      "dev data shape:  (500, 3073)\n",
      "dev labels shape:  (500,)\n"
     ]
    }
   ],
   "source": [
    "def get_CIFAR10_data(num_training=49000, num_validation=1000, num_test=1000, num_dev=500):\n",
    "    \"\"\"\n",
    "    Load the CIFAR-10 dataset from disk and perform preprocessing to prepare\n",
    "    it for the linear classifier. These are the same steps as we used for the\n",
    "    SVM, but condensed to a single function.  \n",
    "    \"\"\"\n",
    "    # Load the raw CIFAR-10 data\n",
    "    cifar10_dir = 'cs231n/datasets/cifar-10-batches-py'\n",
    "    X_train, y_train, X_test, y_test = load_CIFAR10(cifar10_dir)\n",
    "    \n",
    "    # subsample the data\n",
    "    mask = list(range(num_training, num_training + num_validation))\n",
    "    X_val = X_train[mask]\n",
    "    y_val = y_train[mask]\n",
    "    mask = list(range(num_training))\n",
    "    X_train = X_train[mask]\n",
    "    y_train = y_train[mask]\n",
    "    mask = list(range(num_test))\n",
    "    X_test = X_test[mask]\n",
    "    y_test = y_test[mask]\n",
    "    mask = np.random.choice(num_training, num_dev, replace=False)\n",
    "    X_dev = X_train[mask]\n",
    "    y_dev = y_train[mask]\n",
    "    \n",
    "    # Preprocessing: reshape the image data into rows\n",
    "    X_train = np.reshape(X_train, (X_train.shape[0], -1))\n",
    "    X_val = np.reshape(X_val, (X_val.shape[0], -1))\n",
    "    X_test = np.reshape(X_test, (X_test.shape[0], -1))\n",
    "    X_dev = np.reshape(X_dev, (X_dev.shape[0], -1))\n",
    "    \n",
    "    # Normalize the data: subtract the mean image\n",
    "    mean_image = np.mean(X_train, axis = 0)\n",
    "    X_train -= mean_image\n",
    "    X_val -= mean_image\n",
    "    X_test -= mean_image\n",
    "    X_dev -= mean_image\n",
    "    \n",
    "    # add bias dimension and transform into columns\n",
    "    X_train = np.hstack([X_train, np.ones((X_train.shape[0], 1))])\n",
    "    X_val = np.hstack([X_val, np.ones((X_val.shape[0], 1))])\n",
    "    X_test = np.hstack([X_test, np.ones((X_test.shape[0], 1))])\n",
    "    X_dev = np.hstack([X_dev, np.ones((X_dev.shape[0], 1))])\n",
    "    \n",
    "    return X_train, y_train, X_val, y_val, X_test, y_test, X_dev, y_dev\n",
    "\n",
    "\n",
    "# Invoke the above function to get our data.\n",
    "X_train, y_train, X_val, y_val, X_test, y_test, X_dev, y_dev = get_CIFAR10_data()\n",
    "print('Train data shape: ', X_train.shape)\n",
    "print('Train labels shape: ', y_train.shape)\n",
    "print('Validation data shape: ', X_val.shape)\n",
    "print('Validation labels shape: ', y_val.shape)\n",
    "print('Test data shape: ', X_test.shape)\n",
    "print('Test labels shape: ', y_test.shape)\n",
    "print('dev data shape: ', X_dev.shape)\n",
    "print('dev labels shape: ', y_dev.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Softmax Classifier\n",
    "\n",
    "Your code for this section will all be written inside **cs231n/classifiers/softmax.py**. \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "loss: 2.360490\n",
      "sanity check: 2.302585\n"
     ]
    }
   ],
   "source": [
    "# First implement the naive softmax loss function with nested loops.\n",
    "# Open the file cs231n/classifiers/softmax.py and implement the\n",
    "# softmax_loss_naive function.\n",
    "\n",
    "from cs231n.classifiers.softmax import softmax_loss_naive\n",
    "import time\n",
    "\n",
    "# Generate a random softmax weight matrix and use it to compute the loss.\n",
    "W = np.random.randn(3073, 10) * 0.0001\n",
    "loss, grad = softmax_loss_naive(W, X_dev, y_dev, 0.0)\n",
    "\n",
    "# As a rough sanity check, our loss should be something close to -log(0.1).\n",
    "print('loss: %f' % loss)\n",
    "print('sanity check: %f' % (-np.log(0.1)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Inline Question 1:\n",
    "Why do we expect our loss to be close to -log(0.1)? Explain briefly.**\n",
    "\n",
    "**Your answer:** Because with small initial random weights, all probabilities assigned to all classes are equal amongst all classes. Since there are 10 clases, the loss should be -log(1/10) (ie. -log(0.1))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "numerical: -3.913078 analytic: -3.913078, relative error: 7.774156e-09\n",
      "numerical: 2.944017 analytic: 2.944016, relative error: 1.600424e-08\n",
      "numerical: 1.724612 analytic: 1.724612, relative error: 3.076847e-08\n",
      "numerical: 1.488420 analytic: 1.488420, relative error: 6.051570e-08\n",
      "numerical: 1.805924 analytic: 1.805924, relative error: 3.663317e-08\n",
      "numerical: -2.526835 analytic: -2.526835, relative error: 1.536988e-08\n",
      "numerical: 1.121051 analytic: 1.121051, relative error: 2.821812e-09\n",
      "numerical: 0.547604 analytic: 0.547604, relative error: 2.606748e-08\n",
      "numerical: -0.495187 analytic: -0.495187, relative error: 1.386055e-08\n",
      "numerical: 1.166179 analytic: 1.166179, relative error: 5.488790e-08\n",
      "numerical: 0.129210 analytic: 0.129210, relative error: 3.100458e-07\n",
      "numerical: -4.583949 analytic: -4.583949, relative error: 7.675430e-09\n",
      "numerical: 0.427770 analytic: 0.427770, relative error: 2.206212e-07\n",
      "numerical: -5.500289 analytic: -5.500289, relative error: 5.874661e-09\n",
      "numerical: 1.003728 analytic: 1.003728, relative error: 8.692799e-08\n",
      "numerical: 2.997838 analytic: 2.997838, relative error: 4.433544e-08\n",
      "numerical: -2.462659 analytic: -2.462659, relative error: 4.012890e-09\n",
      "numerical: -1.571576 analytic: -1.571576, relative error: 3.729994e-08\n",
      "numerical: 1.962441 analytic: 1.962441, relative error: 8.063131e-09\n",
      "numerical: 0.183956 analytic: 0.183956, relative error: 7.467689e-08\n"
     ]
    }
   ],
   "source": [
    "# Complete the implementation of softmax_loss_naive and implement a (naive)\n",
    "# version of the gradient that uses nested loops.\n",
    "loss, grad = softmax_loss_naive(W, X_dev, y_dev, 0.0)\n",
    "\n",
    "# As we did for the SVM, use numeric gradient checking as a debugging tool.\n",
    "# The numeric gradient should be close to the analytic gradient.\n",
    "from cs231n.gradient_check import grad_check_sparse\n",
    "f = lambda w: softmax_loss_naive(w, X_dev, y_dev, 0.0)[0]\n",
    "grad_numerical = grad_check_sparse(f, W, grad, 10)\n",
    "\n",
    "# similar to SVM case, do another gradient check with regularization\n",
    "loss, grad = softmax_loss_naive(W, X_dev, y_dev, 5e1)\n",
    "f = lambda w: softmax_loss_naive(w, X_dev, y_dev, 5e1)[0]\n",
    "grad_numerical = grad_check_sparse(f, W, grad, 10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "naive loss: 2.360490e+00 computed in 0.128881s\n",
      "vectorized loss: 2.360490e+00 computed in 0.012322s\n",
      "Loss difference: 0.000000\n",
      "Gradient difference: 0.000000\n"
     ]
    }
   ],
   "source": [
    "# Now that we have a naive implementation of the softmax loss function and its gradient,\n",
    "# implement a vectorized version in softmax_loss_vectorized.\n",
    "# The two versions should compute the same results, but the vectorized version should be\n",
    "# much faster.\n",
    "tic = time.time()\n",
    "loss_naive, grad_naive = softmax_loss_naive(W, X_dev, y_dev, 0.000005)\n",
    "toc = time.time()\n",
    "print('naive loss: %e computed in %fs' % (loss_naive, toc - tic))\n",
    "\n",
    "from cs231n.classifiers.softmax import softmax_loss_vectorized\n",
    "tic = time.time()\n",
    "loss_vectorized, grad_vectorized = softmax_loss_vectorized(W, X_dev, y_dev, 0.000005)\n",
    "toc = time.time()\n",
    "print('vectorized loss: %e computed in %fs' % (loss_vectorized, toc - tic))\n",
    "\n",
    "# As we did for the SVM, we use the Frobenius norm to compare the two versions\n",
    "# of the gradient.\n",
    "grad_difference = np.linalg.norm(grad_naive - grad_vectorized, ord='fro')\n",
    "print('Loss difference: %f' % np.abs(loss_naive - loss_vectorized))\n",
    "print('Gradient difference: %f' % grad_difference)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "training accuracy: 0.343592\n",
      "validation accuracy: 0.365000 \n",
      "\n",
      "training accuracy: 0.334184\n",
      "validation accuracy: 0.346000 \n",
      "\n",
      "training accuracy: 0.347653\n",
      "validation accuracy: 0.366000 \n",
      "\n",
      "training accuracy: 0.317755\n",
      "validation accuracy: 0.329000 \n",
      "\n",
      "lr 1.000000e-07 reg 2.500000e+04 train accuracy: 0.343592 val accuracy: 0.365000\n",
      "lr 1.000000e-07 reg 5.000000e+04 train accuracy: 0.334184 val accuracy: 0.346000\n",
      "lr 5.000000e-07 reg 2.500000e+04 train accuracy: 0.347653 val accuracy: 0.366000\n",
      "lr 5.000000e-07 reg 5.000000e+04 train accuracy: 0.317755 val accuracy: 0.329000\n",
      "best validation accuracy achieved during cross-validation: 0.366000\n"
     ]
    }
   ],
   "source": [
    "# Use the validation set to tune hyperparameters (regularization strength and\n",
    "# learning rate). You should experiment with different ranges for the learning\n",
    "# rates and regularization strengths; if you are careful you should be able to\n",
    "# get a classification accuracy of over 0.35 on the validation set.\n",
    "from cs231n.classifiers import Softmax\n",
    "results = {}\n",
    "best_val = -1\n",
    "best_softmax = None\n",
    "learning_rates = [1e-7, 5e-7]\n",
    "regularization_strengths = [2.5e4, 5e4]\n",
    "\n",
    "################################################################################\n",
    "# TODO:                                                                        #\n",
    "# Use the validation set to set the learning rate and regularization strength. #\n",
    "# This should be identical to the validation that you did for the SVM; save    #\n",
    "# the best trained softmax classifer in best_softmax.                          #\n",
    "################################################################################\n",
    "for rate in learning_rates:\n",
    "    for strength in regularization_strengths:\n",
    "        W = np.random.randn(3073, 10) * 0.0001\n",
    "        sm = Softmax()\n",
    "        loss_hist = sm.train(X_train, y_train, learning_rate=rate, reg=strength,\n",
    "                              num_iters=1100, verbose=False)\n",
    "        \n",
    "        y_train_pred = sm.predict(X_train)\n",
    "        y_train_acc = np.mean(y_train == y_train_pred)\n",
    "        print('training accuracy: %f' % (y_train_acc, ))\n",
    "        y_val_pred = sm.predict(X_val)\n",
    "        y_val_acc = np.mean(y_val == y_val_pred)\n",
    "        print('validation accuracy: %f \\n' % (y_val_acc, ))\n",
    "        \n",
    "        results[(rate, strength)] = [y_train_acc, y_val_acc]\n",
    "        \n",
    "        if y_val_acc > best_val:\n",
    "            best_val = y_val_acc\n",
    "            best_softmax= sm\n",
    "\n",
    "################################################################################\n",
    "#                              END OF YOUR CODE                                #\n",
    "################################################################################\n",
    "    \n",
    "# Print out results.\n",
    "for lr, reg in sorted(results):\n",
    "    train_accuracy, val_accuracy = results[(lr, reg)]\n",
    "    print('lr %e reg %e train accuracy: %f val accuracy: %f' % (\n",
    "                lr, reg, train_accuracy, val_accuracy))\n",
    "    \n",
    "print('best validation accuracy achieved during cross-validation: %f' % best_val)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "softmax on raw pixels final test set accuracy: 0.367000\n"
     ]
    }
   ],
   "source": [
    "# evaluate on test set\n",
    "# Evaluate the best softmax on test set\n",
    "y_test_pred = best_softmax.predict(X_test)\n",
    "test_accuracy = np.mean(y_test == y_test_pred)\n",
    "print('softmax on raw pixels final test set accuracy: %f' % (test_accuracy, ))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAlMAAAF8CAYAAADrUz6WAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f0c2cba6ac8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Visualize the learned weights for each class\n",
    "w = best_softmax.W[:-1,:] # strip out the bias\n",
    "w = w.reshape(32, 32, 3, 10)\n",
    "\n",
    "w_min, w_max = np.min(w), np.max(w)\n",
    "\n",
    "classes = ['plane', 'car', 'bird', 'cat', 'deer', 'dog', 'frog', 'horse', 'ship', 'truck']\n",
    "for i in range(10):\n",
    "    plt.subplot(2, 5, i + 1)\n",
    "    \n",
    "    # Rescale the weights to be between 0 and 255\n",
    "    wimg = 255.0 * (w[:, :, :, i].squeeze() - w_min) / (w_max - w_min)\n",
    "    plt.imshow(wimg.astype('uint8'))\n",
    "    plt.axis('off')\n",
    "    plt.title(classes[i])"
   ]
  }
 ],
 "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": "3.5.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}