RBM_Classification.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "%run \"RBM.ipynb\"\n",
    "\n",
    "class RBM_Classification(RBM):\n",
    "    # n_k is the number of classes\n",
    "    def __init__(self, n_x, n_h, n_k, n_d=1, W=None, b=None, c=None, k=1):\n",
    "        self.n_x = n_x\n",
    "        self.n_k = n_k\n",
    "        self.n_d = n_d  \n",
    "        # label is represented with n_k*n_d dots, where label i is the i-th group of n_d dots\n",
    "        # np.zeros(n_k*n_d)[i*n_d:+n_d] = 1\n",
    "        super().__init__(n_x+n_k*n_d, n_h, W, b, c, k)\n",
    "        return\n",
    "        \n",
    "    def convert_input(self, X, Label):\n",
    "        n_sample = X.shape[0]\n",
    "        n_k = self.n_k; n_d = self.n_d; n_x = self.n_x\n",
    "        assert X.shape == (n_sample, n_x)\n",
    "\n",
    "        if Label is not None:\n",
    "            Label = Label.flatten()\n",
    "            #attach label to data\n",
    "            Y = np.zeros(shape=(n_sample, n_k))\n",
    "            Index = np.arange(n_sample)\n",
    "            Y[Index, Label] = 1\n",
    "            Y = Y.repeat(n_d, axis=1)\n",
    "        else:\n",
    "            Y = np.zeros(shape=(n_sample, n_k*n_d))\n",
    "\n",
    "        V = np.append(X, Y, axis=1)\n",
    "        assert V.shape == (n_sample, self.n_v)        \n",
    "        return V\n",
    "\n",
    "    # this is the API for app to use\n",
    "    def train(self, X, Label, n_epoch=1, batch_size=10, learning=0.01, save_file=None):\n",
    "        if self.training_done: return\n",
    "\n",
    "        save_file = save_file + \".rbm_classify\"\n",
    "        V = self.convert_input(X, Label)\n",
    "        self.train_model(V, n_epoch, batch_size, learning, save_file)\n",
    "\n",
    "        self.training_done = True\n",
    "        return\n",
    "\n",
    "    def classify(self, x):\n",
    "        #set_trace()\n",
    "        x = x.reshape(1,-1)\n",
    "        v = self.convert_input(x, None)\n",
    "        vp, vs = super().reconstruct(v)\n",
    "\n",
    "        n_k = self.n_k; n_d = self.n_d; n_x = self.n_x\n",
    "        result = vp[0, n_x:]\n",
    "        result = result.reshape(n_k, -1).sum(axis=1) # collapse the k*d dots into k sum dots.\n",
    "        pred = np.argmax(result)\n",
    "\n",
    "        return vp[0, :n_x], pred\n",
    "    \n",
    "    def create_label_input(self, label, weight):\n",
    "        n_k = self.n_k; n_d = self.n_d\n",
    "        classes = np.zeros(n_k)\n",
    "        classes[label] = weight\n",
    "        classes = classes.repeat(n_d)\n",
    "        return classes\n",
    "\n",
    "    def reconstruct(self, x, label, k_gibbs, init_w, w):\n",
    "        # this is a very simple generation algorithm with single-layer RBM\n",
    "        classes = self.create_label_input(label, init_w)\n",
    "        v = np.append(x, classes)\n",
    "        vp = v.reshape(1,-1)\n",
    "        \n",
    "        classes = self.create_label_input(label, w)\n",
    "        n_x = self.n_x\n",
    "        for i in range(k_gibbs):\n",
    "            vp, vs = super().reconstruct(vp)\n",
    "            vp[0, n_x:] = classes\n",
    "\n",
    "        return vp[0, :n_x]\n",
    "            \n",
    "    def reconstruct_from_label(self, label, k_gibbs, init_w, w):\n",
    "        # this is a very simple generation algorithm with single-layer RBM        \n",
    "        n_x = self.n_x\n",
    "        #np.random.seed(1234)\n",
    "        #x = np.random.binomial(1, 0.2, n_x)\n",
    "        #x = np.zeros(n_x)\n",
    "        #x = np.random.uniform(0, 0.8, n_x)\n",
    "        x = np.full(n_x, 0.1)\n",
    "        v = self.reconstruct(x, label, k_gibbs, init_w, w)\n",
    "        return v\n",
    "\n",
    "    def save_model(self, save_file):\n",
    "        dict = {'n_x':self.n_x, 'n_h':self.n_h, 'n_k':self.n_k, 'n_d':self.n_d, 'W':self.W, 'b':self.b, 'c':self.c}\n",
    "        save_file += \".(\"+ str(self.n_x) + \"+\" + str(self.n_k) + \"x\" + str(self.n_d) + \")x\" + str(self.n_h)\n",
    "        with open(save_file, 'wb') as f:\n",
    "            pickle.dump(dict, f, protocol=pickle.HIGHEST_PROTOCOL)\n",
    "\n",
    "        return\n",
    "\n",
    "    @classmethod\n",
    "    def load_model(cls, load_file):\n",
    "        with open(load_file, 'rb') as f:\n",
    "            m = pickle.load(f)\n",
    "\n",
    "        rbm = cls(m['n_x'], m['n_h'], m['n_k'], m['n_d'], m['W'], m['b'], m['c'])\n",
    "        return rbm"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# testing of the RBM Classification_Generation code above\n",
    "\n",
    "%run \"mnist.ipynb\"\n",
    "import matplotlib as mpl\n",
    "import matplotlib.pyplot as plt\n",
    "from mpl_toolkits.axes_grid1 import Grid\n",
    "import math\n",
    "\n",
    "class MNIST_RBM_CLASSIFY:\n",
    "    def __init__(self, n_x, n_h, n_k=10, n_d=1, load_file=None, folder=\"../convolution-network\"):\n",
    "\n",
    "        if load_file is None:\n",
    "            self.net= RBM_Classification(n_x, n_h, n_k, n_d)\n",
    "        else:\n",
    "            self.net = RBM_Classification.load_model(load_file)\n",
    "            \n",
    "        self.train_input = MnistInput(\"train\", folder)\n",
    "        self.test_input = MnistInput(\"test\", folder)\n",
    "\n",
    "        return\n",
    "\n",
    "    def train(self, train_size=-1, n_epoch=100, batch_size=10, learning=0.1):\n",
    "        if self.net.training_done: return\n",
    "        \n",
    "        X = []; Y=[]\n",
    "        n_x = 0\n",
    "        for x, y in self.train_input.read(train_size):\n",
    "            X.append(x)\n",
    "            Y.append(y)\n",
    "            n_x += 1\n",
    "\n",
    "        X = (np.array(X).reshape(n_x, -1) > 30)*1\n",
    "        Label = np.array(Y).reshape(n_x, -1)\n",
    "        \n",
    "        self.net.train(X, Label, n_epoch, batch_size, learning, save_file=\"mnist\")\n",
    "        return\n",
    "\n",
    "    def classify(self, test_size=-1, output_size=20):\n",
    "        n_total=0; n_correct = 0; n_output = 0\n",
    "        X=[]; Y=[]; Recon=[]; Preds=[]\n",
    "        for x, y in self.test_input.read(test_size):\n",
    "            n_total += 1\n",
    "            reco, pred = self.net.classify(x/255)\n",
    "            if y == pred: n_correct += 1 \n",
    "            if n_output < output_size:\n",
    "                Preds.append(pred)\n",
    "                Recon.append(reco)\n",
    "                X.append(x)\n",
    "                Y.append(y)\n",
    "                n_output += 1\n",
    "\n",
    "        accuracy = n_correct/n_total\n",
    "        print(\"Classification accuracy: {}\".format(accuracy))\n",
    "\n",
    "        # output\n",
    "        d = 28\n",
    "        ncols = 10\n",
    "        nrows = int(output_size/5)\n",
    "        fig = plt.figure(figsize=(ncols, int(nrows*2)), dpi=100)\n",
    "        grid = Grid(fig, rect=111, nrows_ncols=(nrows,ncols))\n",
    "\n",
    "        for i, ax in enumerate(grid):\n",
    "            j = i//2\n",
    "            if i%2 == 0:\n",
    "                ax.imshow(X[j].reshape(d,d), cmap=mpl.cm.Greys)\n",
    "                ax.set_title(\"Orig: {}\".format(Y[j]), y=0)\n",
    "            else:\n",
    "                ax.imshow(Recon[j].reshape(d,d), cmap=mpl.cm.Greys)\n",
    "                ax.set_title(\"Reco: {}\".format(Preds[j]), y=0)\n",
    "\n",
    "            ax.set_axis_off()\n",
    "\n",
    "        fig.text(0.5,1, \"Classified images and reconstructed digits\", fontsize=20, horizontalalignment='center')\n",
    "        fig.tight_layout()\n",
    "        plt.show()        \n",
    "        return\n",
    "    \n",
    "    def generate(self, k_gibbs=50, init_w=30, w=30, n_k=10):\n",
    "        # 50,30,30 is for mnist_rbm_classify.(784+10)x85.epochs100\n",
    "        digits = []\n",
    "        for i in range( n_k ):\n",
    "            digit = self.net.reconstruct_from_label(i, k_gibbs, init_w, w)\n",
    "            digits.append(digit)\n",
    "        \n",
    "        # output\n",
    "        d = 28\n",
    "        ncols = n_k\n",
    "        nrows = 1\n",
    "        fig = plt.figure(figsize=(ncols, int(nrows*2)), dpi=100)\n",
    "        grid = Grid(fig, rect=111, nrows_ncols=(nrows,ncols))\n",
    "\n",
    "        for i, ax in enumerate(grid):\n",
    "            ax.imshow(digits[i].reshape(d,d), cmap=mpl.cm.Greys)\n",
    "            ax.set_title(i)\n",
    "            ax.set_axis_off()\n",
    "\n",
    "        fig.suptitle('Generated digit images from numbers')\n",
    "        fig.tight_layout()\n",
    "        fig.subplots_adjust(top=0.7)    \n",
    "        plt.show()        \n",
    "        return\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "run_control": {
     "marked": false
    },
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA9sAAADICAYAAAD4I/J0AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAACAmtHZBgAAAACgZv8PTHFXuqY8i+sAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x104a83080>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "mnist = None\n",
    "if __name__ == \"__main__\" and '__file__' not in globals():\n",
    "    np.seterr(all='raise')\n",
    "    mnist = MNIST_RBM_CLASSIFY(28*28, n_h=28*3+1, load_file=\"trained_models/mnist_rbm_classify.epochs10.(784+10)x85\")\n",
    "    #mnist = MNIST_RBM_CLASSIFY(28*28, n_h=28*20+1)\n",
    "    #mnist.train(train_size=1000, n_epoch=2, batch_size=10, learning=0.1)\n",
    "    plt.close('all')\n",
    "    #mnist.classify()\n",
    "    mnist.generate(20,50,8)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "run_control": {
     "marked": false
    },
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Classification accuracy: 0.9166\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA94AAAM7CAYAAAC85EgaAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\njYhyAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x10756ee80>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "mnist = None\n",
    "if __name__ == \"__main__\" and '__file__' not in globals():\n",
    "    np.seterr(all='raise')\n",
    "    mnist = MNIST_RBM_CLASSIFY(28*28, n_h=28*20+1, load_file=\"trained_models/mnist_rbm_classify.(784+10)x561.epochs100\")\n",
    "    #mnist = MNIST_RBM_CLASSIFY(28*28, n_h=28*20+1)\n",
    "    mnist.train(train_size=1000, n_epoch=2, batch_size=10, learning=0.1)\n",
    "    plt.close('all')\n",
    "    mnist.classify()\n",
    "    #mnist.generate(10,5,5)"
   ]
  }
 ],
 "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": "3.6.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}