assignment1/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": 1,
   "metadata": {},
   "outputs": [],
   "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": 2,
   "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": "markdown",
   "metadata": {},
   "source": [
    "$$\n",
    "L_i = -f_{y_i} + log \\sum_j e^{f_j}\n",
    "$$"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "loss: 2.398304\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:** \n",
    "\n",
    "\n",
    "Because initially $X$ and $W$ have mean = 0 inner categorical distribution (softmax function) has equal probability for each class.\n",
    "Thus:\n",
    "$$\n",
    "mean(L_i) = mean(-log\\frac{e^{f_j}}{\\sum_j e^{y_j}}) = - log(1/number\\_of\\_classes) = - log(0.1)\n",
    "$$"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "numerical: 1.617808 analytic: 1.617808, relative error: 4.691552e-08\n",
      "numerical: 3.409211 analytic: 3.409211, relative error: 7.454114e-09\n",
      "numerical: -2.391061 analytic: -2.391061, relative error: 2.795177e-08\n",
      "numerical: 0.360223 analytic: 0.360223, relative error: 1.011852e-09\n",
      "numerical: 1.191565 analytic: 1.191565, relative error: 1.003437e-08\n",
      "numerical: 0.147612 analytic: 0.147612, relative error: 2.254818e-07\n",
      "numerical: 0.219801 analytic: 0.219801, relative error: 1.246302e-07\n",
      "numerical: -2.279283 analytic: -2.279283, relative error: 5.509348e-09\n",
      "numerical: 0.652616 analytic: 0.652616, relative error: 5.786774e-08\n",
      "numerical: -0.958676 analytic: -0.958676, relative error: 1.331018e-08\n",
      "numerical: 1.461864 analytic: 1.461864, relative error: 5.432430e-08\n",
      "numerical: 1.611650 analytic: 1.611650, relative error: 3.635080e-08\n",
      "numerical: -1.683001 analytic: -1.683001, relative error: 1.169177e-08\n",
      "numerical: 1.635636 analytic: 1.635635, relative error: 5.505049e-08\n",
      "numerical: 2.608148 analytic: 2.608148, relative error: 1.224492e-08\n",
      "numerical: 0.785262 analytic: 0.785262, relative error: 4.856924e-08\n",
      "numerical: -1.876761 analytic: -1.876762, relative error: 2.544329e-08\n",
      "numerical: 3.802986 analytic: 3.802986, relative error: 1.896227e-08\n",
      "numerical: -0.478247 analytic: -0.478247, relative error: 4.664491e-08\n",
      "numerical: 1.741047 analytic: 1.741047, relative error: 3.792013e-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": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "naive loss: 2.398304e+00 computed in 0.128395s\n",
      "vectorized loss: 2.398304e+00 computed in 0.012475s\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": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "#\n",
      "learning_rate 1e-07\n",
      "regularization_strength 25000.0\n",
      "#\n",
      "iteration 0 / 1500: loss 771.249091\n",
      "iteration 100 / 1500: loss 282.698114\n",
      "iteration 200 / 1500: loss 105.473012\n",
      "iteration 300 / 1500: loss 39.540603\n",
      "iteration 400 / 1500: loss 15.807770\n",
      "iteration 500 / 1500: loss 6.823686\n",
      "iteration 600 / 1500: loss 3.937375\n",
      "iteration 700 / 1500: loss 2.871500\n",
      "iteration 800 / 1500: loss 2.107952\n",
      "iteration 900 / 1500: loss 2.100126\n",
      "iteration 1000 / 1500: loss 1.777365\n",
      "iteration 1100 / 1500: loss 2.002794\n",
      "iteration 1200 / 1500: loss 1.937557\n",
      "iteration 1300 / 1500: loss 1.727722\n",
      "iteration 1400 / 1500: loss 1.899078\n",
      "\n",
      "#\n",
      "learning_rate 1e-07\n",
      "regularization_strength 50000.0\n",
      "#\n",
      "iteration 0 / 1500: loss 1560.218610\n",
      "iteration 100 / 1500: loss 209.886539\n",
      "iteration 200 / 1500: loss 30.031017\n",
      "iteration 300 / 1500: loss 5.794698\n",
      "iteration 400 / 1500: loss 2.631693\n",
      "iteration 500 / 1500: loss 2.213578\n",
      "iteration 600 / 1500: loss 2.096099\n",
      "iteration 700 / 1500: loss 2.043242\n",
      "iteration 800 / 1500: loss 2.243055\n",
      "iteration 900 / 1500: loss 2.044673\n",
      "iteration 1000 / 1500: loss 2.331667\n",
      "iteration 1100 / 1500: loss 2.132263\n",
      "iteration 1200 / 1500: loss 2.002235\n",
      "iteration 1300 / 1500: loss 2.163395\n",
      "iteration 1400 / 1500: loss 2.174670\n",
      "\n",
      "#\n",
      "learning_rate 5e-07\n",
      "regularization_strength 25000.0\n",
      "#\n",
      "iteration 0 / 1500: loss 780.012363\n",
      "iteration 100 / 1500: loss 7.171205\n",
      "iteration 200 / 1500: loss 2.158529\n",
      "iteration 300 / 1500: loss 2.175014\n",
      "iteration 400 / 1500: loss 2.133895\n",
      "iteration 500 / 1500: loss 2.037722\n",
      "iteration 600 / 1500: loss 2.524336\n",
      "iteration 700 / 1500: loss 2.415645\n",
      "iteration 800 / 1500: loss 2.262346\n",
      "iteration 900 / 1500: loss 2.252650\n",
      "iteration 1000 / 1500: loss 2.218206\n",
      "iteration 1100 / 1500: loss 2.323351\n",
      "iteration 1200 / 1500: loss 2.192917\n",
      "iteration 1300 / 1500: loss 2.366462\n",
      "iteration 1400 / 1500: loss 2.167075\n",
      "\n",
      "#\n",
      "learning_rate 5e-07\n",
      "regularization_strength 50000.0\n",
      "#\n",
      "iteration 0 / 1500: loss 1538.378816\n",
      "iteration 100 / 1500: loss 2.086087\n",
      "iteration 200 / 1500: loss 2.269833\n",
      "iteration 300 / 1500: loss 2.481761\n",
      "iteration 400 / 1500: loss 2.180135\n",
      "iteration 500 / 1500: loss 2.303638\n",
      "iteration 600 / 1500: loss 2.254269\n",
      "iteration 700 / 1500: loss 2.399879\n",
      "iteration 800 / 1500: loss 2.193813\n",
      "iteration 900 / 1500: loss 2.513852\n",
      "iteration 1000 / 1500: loss 2.395803\n",
      "iteration 1100 / 1500: loss 2.683795\n",
      "iteration 1200 / 1500: loss 2.218959\n",
      "iteration 1300 / 1500: loss 2.282853\n",
      "iteration 1400 / 1500: loss 2.536470\n",
      "lr 1.000000e-07 reg 2.500000e+04 train accuracy: 0.318122 val accuracy: 0.321000\n",
      "lr 1.000000e-07 reg 5.000000e+04 train accuracy: 0.288980 val accuracy: 0.307000\n",
      "lr 5.000000e-07 reg 2.500000e+04 train accuracy: 0.262694 val accuracy: 0.281000\n",
      "lr 5.000000e-07 reg 5.000000e+04 train accuracy: 0.211653 val accuracy: 0.217000\n",
      "best validation accuracy achieved during cross-validation: 0.321000\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 learning_rate in learning_rates:\n",
    "    for regularization_strength in regularization_strengths:\n",
    "        print('\\n#')\n",
    "        print('learning_rate', learning_rate)\n",
    "        print('regularization_strength', regularization_strength)\n",
    "        print('#')\n",
    "        classifier = Softmax()\n",
    "        classifier.train(X_train, y_train, \n",
    "                         learning_rate=learning_rate, \n",
    "                         reg=regularization_strength,\n",
    "                         num_iters=1500, \n",
    "                         verbose=True,\n",
    "                        )\n",
    "        trained_accuracy = np.mean(y_train == classifier.predict(X_train))\n",
    "        val_accuracy = np.mean(y_val == classifier.predict(X_val))        \n",
    "        results[(\n",
    "            learning_rate, regularization_strength,\n",
    "        )] = (\n",
    "            trained_accuracy, \n",
    "            val_accuracy,\n",
    "        )\n",
    "        if best_val < val_accuracy:\n",
    "            best_val = val_accuracy\n",
    "            best_softmax = classifier\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": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "softmax on raw pixels final test set accuracy: 0.315000\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": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAlMAAAF8CAYAAADrUz6WAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fcde7d2d240>"
      ]
     },
     "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": "3.6.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}