RBFN_MNIST.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 08_NN_MNIST"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-09-16T13:20:30.345325Z",
     "start_time": "2017-09-16T13:20:29.860841Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<torch._C.Generator at 0x7fdab404c390>"
      ]
     },
     "execution_count": 1,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Pytorch Library\n",
    "import torch\n",
    "import torch.nn.init\n",
    "import torch.nn as nn\n",
    "from torch.autograd import Variable\n",
    "\n",
    "torch.manual_seed(777)  # reproducibility"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-09-16T13:20:30.414079Z",
     "start_time": "2017-09-16T13:20:30.347965Z"
    },
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import torchvision.utils as utils\n",
    "import torchvision.datasets as dsets\n",
    "import torchvision.transforms as transforms"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-09-16T13:20:30.725501Z",
     "start_time": "2017-09-16T13:20:30.491419Z"
    },
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Other Python Library\n",
    "%matplotlib inline\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import random"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Loading MNIST dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-09-16T13:20:31.247585Z",
     "start_time": "2017-09-16T13:20:31.208482Z"
    },
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# MNIST dataset\n",
    "mnist_train = dsets.MNIST(root='data/',\n",
    "                          train=True,\n",
    "                          transform=transforms.ToTensor(),\n",
    "                          download=True)\n",
    "\n",
    "mnist_test = dsets.MNIST(root='data/',\n",
    "                         train=False,\n",
    "                         transform=transforms.ToTensor(),\n",
    "                         download=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-09-16T13:20:32.112067Z",
     "start_time": "2017-09-16T13:20:31.901653Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "torch.Size([60000, 28, 28])\n",
      "torch.Size([60000])\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.text.Text at 0x7fdac0285f28>"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAP8AAAEICAYAAACQ6CLfAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fdac02aa908>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# plot one example\n",
    "print(mnist_train.train_data.size())                 # (60000, 28, 28)\n",
    "print(mnist_train.train_labels.size())               # (60000)\n",
    "\n",
    "idx = 0\n",
    "plt.imshow(mnist_train.train_data[idx,:,:].numpy(), cmap='gray')\n",
    "plt.title('%i' % mnist_train.train_labels[idx])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Data Loader (= 전체 Data를 쪼개서 Batch Size 단위로 불러 오는 역할)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-09-16T13:20:37.941959Z",
     "start_time": "2017-09-16T13:20:37.935016Z"
    },
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Hyper-parameters\n",
    "batch_size = 3000\n",
    "# dataset loader\n",
    "data_loader = torch.utils.data.DataLoader(dataset=mnist_train,\n",
    "                                          batch_size=batch_size,\n",
    "                                          shuffle=True,\n",
    "                                          num_workers=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-09-16T13:20:39.140814Z",
     "start_time": "2017-09-16T13:20:38.405482Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "torch.Size([3000, 1, 28, 28])\n",
      "torch.Size([3000])\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAEIAAAD8CAYAAADDlHLtAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fdac0270470>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "array([4, 1, 6, ..., 4, 0, 5])"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "def imshow(img):\n",
    "    img = img / 2 + 0.5     # unnormalize\n",
    "    npimg = img.numpy()\n",
    "    plt.imshow(np.transpose(npimg, (1, 2, 0)))\n",
    "    plt.show()\n",
    "\n",
    "# #예를 들어서.. data_loader로 부터 데이터 뭉치를 하나만 불러 오자\n",
    "batch_images, batch_labels = next(iter(data_loader))\n",
    "\n",
    "print(batch_images.size())\n",
    "print(batch_labels.size())\n",
    "\n",
    "# show images and print labels\n",
    "imshow(utils.make_grid(batch_images))\n",
    "batch_labels.numpy()\n",
    "\n",
    "# 실제 사용 예시 for loop를 이용해서 image와 label을 읽어 온다.\n",
    "# for batch_images, batch_labels in data_loader:\n",
    "#     print(img.size())\n",
    "#     print(label)\n",
    "\n",
    "#enumerate를 사용하는 경우 image와 label을 읽는 것 뿐만 아니라 iter도 같이 계산\n",
    "# for i, (batch_images, batch_labels) in enumerate(data_loader):\n",
    "#     print(batch_images.size())\n",
    "#     print(batch_labels)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-09-16T13:20:39.149321Z",
     "start_time": "2017-09-16T13:20:39.142864Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "8.3733829179633457"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "type(batch_images)\n",
    "import numpy as np\n",
    "np.sqrt((batch_images[0,] - batch_images[6,] ).pow(2).sum())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Define Neural Network Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-09-16T13:20:42.563548Z",
     "start_time": "2017-09-16T13:20:39.902764Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "RbfNet (\n",
      "  (linear): Linear (3000 -> 10)\n",
      ")\n"
     ]
    }
   ],
   "source": [
    "class RbfNet(nn.Module):\n",
    "    def __init__(self, centers, num_class=10):\n",
    "        super(RbfNet, self).__init__()\n",
    "        self.centers = centers\n",
    "        self.num_centers = centers.size(0)\n",
    "        self.num_class = num_class\n",
    "    \n",
    "        self.linear = torch.nn.Linear(self.num_centers, self.num_class, bias=True)\n",
    "        self.beta = nn.Parameter(torch.ones(1,self.num_centers)/10)\n",
    "\n",
    "    def radial_fun(self, batches):\n",
    "        n_input = batches.size(0) # number of inputs\n",
    "        A = self.centers.view(self.num_centers,-1).repeat(n_input,1,1)\n",
    "        B = batches.view(n_input,-1).unsqueeze(1).repeat(1,self.num_centers,1)\n",
    "        C = torch.exp(-self.beta.mul((A-B).pow(2).sum(2,keepdim=False).sqrt() ) )\n",
    "        return C\n",
    "    \n",
    "    def forward(self, batches):\n",
    "        radial_val = self.radial_fun(batches)\n",
    "        class_score = self.linear(radial_val)\n",
    "        return class_score\n",
    "\n",
    "\n",
    "centers = batch_images\n",
    "#centers = torch.rand(1000,28*28)\n",
    "model = RbfNet(Variable(centers.cuda()), num_class=10)\n",
    " \n",
    "model.cuda()\n",
    "\n",
    "#model.load_state_dict(torch.load('NN.pkl'))  # Load the Trained Model\n",
    "print(model)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-09-16T13:20:42.573020Z",
     "start_time": "2017-09-16T13:20:42.566844Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "torch.Size([3000, 1, 28, 28])"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "centers.size()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-09-16T13:20:43.368876Z",
     "start_time": "2017-09-16T13:20:43.356931Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Parameter containing:\n",
      " 0.1000  0.1000  0.1000  ...   0.1000  0.1000  0.1000\n",
      "[torch.cuda.FloatTensor of size 1x3000 (GPU 0)]\n",
      "\n",
      "Parameter containing:\n",
      " 6.1721e-03 -1.0911e-02 -1.2007e-02  ...   1.9036e-03  1.7342e-02  9.8540e-03\n",
      "-1.0860e-02  1.7958e-02 -6.9733e-03  ...  -1.2285e-02  1.1715e-02 -1.3497e-02\n",
      " 1.0219e-02 -3.8969e-03 -9.9665e-04  ...   1.5645e-03  6.1913e-03  1.3493e-02\n",
      "                ...                   ⋱                   ...                \n",
      "-6.1587e-03  1.0285e-02 -1.4230e-02  ...  -8.5106e-03 -1.5129e-02  5.6613e-03\n",
      "-1.7507e-02 -1.3444e-02 -1.2294e-02  ...  -3.7234e-03  9.8994e-03  1.6329e-02\n",
      "-1.5913e-02  8.7634e-03 -1.8199e-02  ...  -1.2718e-02  1.2614e-02  3.5766e-03\n",
      "[torch.cuda.FloatTensor of size 10x3000 (GPU 0)]\n",
      "\n",
      "Parameter containing:\n",
      "1.00000e-02 *\n",
      "  1.0994\n",
      " -1.6152\n",
      "  0.8603\n",
      "  1.7463\n",
      " -0.1176\n",
      "  0.7831\n",
      "  0.2706\n",
      " -1.6637\n",
      "  0.8689\n",
      "  1.4076\n",
      "[torch.cuda.FloatTensor of size 10 (GPU 0)]\n",
      "\n"
     ]
    }
   ],
   "source": [
    "for parameter in model.parameters():\n",
    "    print(parameter) \n",
    "    \n",
    "#model.load_state_dict(torch.load('NN.pkl')) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-09-16T13:20:43.580718Z",
     "start_time": "2017-09-16T13:20:43.574127Z"
    },
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Hyper-parameters\n",
    "batch_size = 150\n",
    "# dataset loader\n",
    "data_loader = torch.utils.data.DataLoader(dataset=mnist_train,\n",
    "                                          batch_size=batch_size,\n",
    "                                          shuffle=True,\n",
    "                                          num_workers=4)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-09-16T13:20:44.318770Z",
     "start_time": "2017-09-16T13:20:43.822982Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Variable containing:\n",
       "-0.0360  0.2700 -0.2152  ...   0.1751 -0.0054  0.1743\n",
       "-0.0227  0.2793 -0.2067  ...   0.1456  0.0239  0.1143\n",
       "-0.0374  0.2703 -0.2241  ...   0.1452  0.0258  0.1127\n",
       "          ...             ⋱             ...          \n",
       "-0.0093  0.2430 -0.1860  ...   0.1443  0.0357  0.0831\n",
       "-0.0116  0.2663 -0.1878  ...   0.1160  0.0358  0.1098\n",
       "-0.0253  0.2802 -0.1791  ...   0.1362 -0.0096  0.1107\n",
       "[torch.cuda.FloatTensor of size 150x10 (GPU 0)]"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#for _ in range(10):\n",
    "batch_images, batch_labels = next(iter(data_loader))\n",
    "model(Variable(batch_images.cuda(), requires_grad=False))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Cost 함수 정의"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-09-16T13:20:44.527175Z",
     "start_time": "2017-09-16T13:20:44.522233Z"
    },
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Softmax 함수가 Cost를 계산할 때 내장되어 있다.\n",
    "cost_func = torch.nn.CrossEntropyLoss()    "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Model 학습"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "start_time": "2017-09-16T13:24:44.489Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[Epoch:    1] cost = 1.20566511\n",
      "[Epoch:    2] cost = 0.511912048\n",
      "[Epoch:    3] cost = 0.442301959\n",
      "[Epoch:    4] cost = 0.403429091\n",
      "[Epoch:    5] cost = 0.379295021\n",
      "[Epoch:    6] cost = 0.358320385\n",
      "[Epoch:    7] cost = 0.337369591\n",
      "[Epoch:    8] cost = 0.329559326\n",
      "[Epoch:    9] cost = 0.324487895\n",
      "[Epoch:   10] cost = 0.305682033\n",
      "[Epoch:   11] cost = 0.301581651\n",
      "[Epoch:   12] cost = 0.285979122\n",
      "[Epoch:   13] cost = 0.287023187\n",
      "[Epoch:   14] cost = 0.274200559\n",
      "[Epoch:   15] cost = 0.266997248\n",
      "[Epoch:   16] cost = 0.26618138\n",
      "[Epoch:   17] cost = 0.267345041\n",
      "[Epoch:   18] cost = 0.254716784\n",
      "[Epoch:   19] cost = 0.243363231\n",
      "[Epoch:   20] cost = 0.24887237\n",
      "[Epoch:   21] cost = 0.241127446\n",
      "[Epoch:   22] cost = 0.2332986\n",
      "[Epoch:   23] cost = 0.235295102\n",
      "[Epoch:   24] cost = 0.224552915\n",
      "[Epoch:   25] cost = 0.226922065\n",
      "[Epoch:   26] cost = 0.222697884\n",
      "[Epoch:   27] cost = 0.222331896\n",
      "[Epoch:   28] cost = 0.214291126\n",
      "[Epoch:   29] cost = 0.21221523\n",
      "[Epoch:   30] cost = 0.212063432\n",
      "[Epoch:   31] cost = 0.210365966\n",
      "[Epoch:   32] cost = 0.210659832\n",
      "[Epoch:   33] cost = 0.20967555\n",
      "[Epoch:   34] cost = 0.200133458\n",
      "[Epoch:   35] cost = 0.201444954\n",
      "[Epoch:   36] cost = 0.196319625\n",
      "[Epoch:   37] cost = 0.202814713\n",
      "[Epoch:   38] cost = 0.19189702\n",
      "[Epoch:   39] cost = 0.194286183\n",
      "[Epoch:   40] cost = 0.19041051\n",
      "[Epoch:   41] cost = 0.190453455\n"
     ]
    }
   ],
   "source": [
    "# Hyper-parameters\n",
    "learning_rate = 0.01 \n",
    "training_epochs = 100\n",
    "\n",
    "# Adam Optimizer\n",
    "optimizer = torch.optim.Adam(model.parameters(), lr=learning_rate)\n",
    "\n",
    "# Train model\n",
    "for epoch in range(training_epochs):\n",
    "    avg_cost = 0\n",
    "    total_batch = len(mnist_train) // batch_size\n",
    "\n",
    "    for i, (batch_images, batch_labels) in enumerate(data_loader):\n",
    "        \n",
    "        # 이미지를 [batch_size x 784] size 행렬로 변환\n",
    "        \n",
    "        X = Variable(batch_images.view(-1, 28 * 28)).cuda()\n",
    "        Y = Variable(batch_labels).cuda()        # label is not one-hot encoded\n",
    "\n",
    "        optimizer.zero_grad()             # Zero Gradient Container\n",
    "        Y_prediction = model(X)           # Forward Propagation\n",
    "        cost = cost_func(Y_prediction, Y) # compute cost\n",
    "        cost.backward()                   # compute gradient\n",
    "        optimizer.step()                  # gradient update\n",
    "\n",
    "        avg_cost += cost / total_batch\n",
    "\n",
    "    print(\"[Epoch: {:>4}] cost = {:>.9}\".format(epoch + 1, avg_cost.data[0]))\n",
    "\n",
    "print('Learning Finished!')\n",
    "torch.save(model.state_dict(), 'NN.pkl')  # Save the Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-08-07T06:40:06.003112Z",
     "start_time": "2017-08-07T06:41:33.928Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "OrderedDict([('beta', \n",
       "               0.2166  0.2861  0.2769  ...   0.2473  0.1988  0.1221\n",
       "              [torch.cuda.FloatTensor of size 1x5000 (GPU 0)]),\n",
       "             ('linear.weight', \n",
       "              -3.7569e-02 -3.6120e-01  3.8630e-01  ...   3.5091e-01 -2.1983e-01  7.4422e-02\n",
       "              -2.9959e-01  6.0009e-01 -3.6100e-01  ...  -2.9455e-01 -5.9414e-02  1.2984e-01\n",
       "               2.1883e-01 -3.4451e-01  6.8182e-02  ...  -1.0338e-01 -1.4275e-01 -1.1496e-01\n",
       "                              ...                   ⋱                   ...                \n",
       "              -6.3282e-02  6.9182e-02 -3.1164e-01  ...  -2.9166e-01  4.3582e-02  3.0526e-01\n",
       "              -2.7239e-02 -5.3782e-02 -1.2934e-01  ...  -1.1819e-01 -1.2814e-01 -1.8660e-01\n",
       "              -1.2194e-01  6.0939e-02 -2.4678e-01  ...  -4.3307e-01 -2.3870e-01 -5.5325e-02\n",
       "              [torch.cuda.FloatTensor of size 10x5000 (GPU 0)]),\n",
       "             ('linear.bias', \n",
       "               0.0258\n",
       "               0.3024\n",
       "              -0.0371\n",
       "               0.0442\n",
       "              -0.0058\n",
       "               0.1398\n",
       "              -0.0591\n",
       "               0.1860\n",
       "              -0.3369\n",
       "              -0.0770\n",
       "              [torch.cuda.FloatTensor of size 10 (GPU 0)])])"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.state_dict()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Test Dataset 이용한 모형 성능 측정"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 139,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-08-07T06:40:06.004090Z",
     "start_time": "2017-08-07T06:41:35.101Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy of the network on the 10000 test images: 53 %\n"
     ]
    }
   ],
   "source": [
    "# Test the Model\n",
    "correct = 0\n",
    "total = 0\n",
    "for images, labels in mnist_test:\n",
    "    images = Variable(images.view(-1, 28*28)).cuda()\n",
    "    outputs = model(images)\n",
    "    _, predicted = torch.max(outputs.data, 1)\n",
    "    total += 1\n",
    "    correct += (predicted == labels).sum()\n",
    "\n",
    "print('Accuracy of the network on the 10000 test images: %d %%' % (100 * correct / total))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Random Sample Test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 126,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-08-07T06:40:06.004925Z",
     "start_time": "2017-08-07T06:41:36.197Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Label:  \n",
      " 9\n",
      "[torch.cuda.LongTensor of size 1 (GPU 0)]\n",
      "\n",
      "Prediction:  \n",
      " 9\n",
      "[torch.cuda.LongTensor of size 1 (GPU 0)]\n",
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAP8AAAD8CAYAAAC4nHJkAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAC/pJREFUeJzt3W+IZXd9x/H3t1GfRB8klQ5LXFwroSB5EGUIBUN3B43E\nIGx8EtwHZUuD4wMDFXzQEB9khlIIpVp8JIxkcVNsVEhCFpGqXWY2LRTJJqT5W02UFXfZ7BoiGB/Z\nJN8+uGdl3Mzcc/eec++5s9/3C4a59/zuPee7h/3M75z7O+f+IjORVM+fDF2ApGEYfqkowy8VZfil\nogy/VJThl4oy/FJRhl8qyvBLRb1rnhuLCC8nlGYsM2OS13Xq+SPi9oj4aUS8EhH3dlmXpPmKaa/t\nj4hrgJ8BtwFngSeBI5n54pj32PNLMzaPnv8W4JXM/EVm/h74DnC4w/okzVGX8N8A/Grb87PNsj8S\nEasRcToiTnfYlqSezfwDv8zcADbAw35pkXTp+c8B+7c9/0CzTNIe0CX8TwI3RsSHIuI9wOeAE/2U\nJWnWpj7sz8w3I+Ie4IfANcCxzHyht8okzdTUQ31Tbcxzfmnm5nKRj6S9y/BLRRl+qSjDLxVl+KWi\nDL9UlOGXijL8UlGGXyrK8EtFGX6pKMMvFWX4paIMv1SU4ZeKMvxSUYZfKsrwS0UZfqkowy8VZfil\nogy/VJThl4oy/FJRhl8qyvBLRRl+qSjDLxVl+KWipp6iGyAizgBvAG8Bb2bmch9FaX4OHTo0tv3+\n++/v9P5xVlZWxrZvbW1NvW616xT+xkpmvtbDeiTNkYf9UlFdw5/AjyLiqYhY7aMgSfPR9bD/1sw8\nFxF/Bvw4Iv43M5/Y/oLmj4J/GKQF06nnz8xzze+LwGPALTu8ZiMzl/0wUFosU4c/Iq6NiPddegx8\nCni+r8IkzVaXw/4l4LGIuLSef8vMf++lKkkzF5k5v41FzG9jV5G1tbWx7QcPHty1rcs4/NCajkVX\nKDMn2nEO9UlFGX6pKMMvFWX4paIMv1SU4ZeK6uOuPnXUNpTXdlvt1aptmNJbfrux55eKMvxSUYZf\nKsrwS0UZfqkowy8VZfiloryltwddb5vd3Nwc2942nt1l+23rPnXq1Nj2cbcTQ7favKV3Ot7SK2ks\nwy8VZfilogy/VJThl4oy/FJRhl8qyvv5e9B1muuu6x9nfX19bHvbdwl0NW79bdcIaLbs+aWiDL9U\nlOGXijL8UlGGXyrK8EtFGX6pqNb7+SPiGPAZ4GJm3tQsux74LnAAOAPclZm/ad3YVXo//zy/E2En\n48byZz2OP0uzrH0v75c2fd7P/y3g9suW3QuczMwbgZPNc0l7SGv4M/MJ4PXLFh8GjjePjwN39lyX\npBmb9px/KTPPN49fBZZ6qkfSnHS+tj8zc9y5fESsAqtdtyOpX9P2/BciYh9A8/vibi/MzI3MXM7M\n5Sm3JWkGpg3/CeBo8/go8Hg/5Uial9bwR8TDwH8DfxERZyPibuAB4LaIeBn4ZPNc0h7Ses6fmUd2\nafpEz7UstCHH8ldWVsa279V56tvG2rt+D0IXV/N1AJd4hZ9UlOGXijL8UlGGXyrK8EtFGX6pKL+6\newG0fb32Xh3K02Kz55eKMvxSUYZfKsrwS0UZfqkowy8VZfilohznXwB7+fbRLtOTd5l6XN3Z80tF\nGX6pKMMvFWX4paIMv1SU4ZeKMvxSUY7zN4Ycc24b529rH1d713/XwYMHp962Fps9v1SU4ZeKMvxS\nUYZfKsrwS0UZfqkowy8V1TrOHxHHgM8AFzPzpmbZGvB54NfNy+7LzB/MqsirXdtU1ENOVa2r1yQ9\n/7eA23dY/i+ZeXPzY/ClPaY1/Jn5BPD6HGqRNEddzvnviYhnI+JYRFzXW0WS5mLa8H8D+DBwM3Ae\n+OpuL4yI1Yg4HRGnp9yWpBmYKvyZeSEz38rMt4FvAreMee1GZi5n5vK0RUrq31Thj4h9255+Fni+\nn3IkzcskQ30PA4eA90fEWeB+4FBE3AwkcAb4wgxrlDQDreHPzCM7LH5wBrUMamtra+p272nXXuQV\nflJRhl8qyvBLRRl+qSjDLxVl+KWi/OruCa2srOzatrm5Ofa9e3kocH19fWx7lyHSq3m/7QX2/FJR\nhl8qyvBLRRl+qSjDLxVl+KWiDL9UlOP8PRh3DcA8jBsPbxuHH9KpU6fGts9ynL9t6vEK7Pmlogy/\nVJThl4oy/FJRhl8qyvBLRRl+qSjH+a8CizyWr8Vlzy8VZfilogy/VJThl4oy/FJRhl8qyvBLRbWO\n80fEfuAhYAlIYCMzvx4R1wPfBQ4AZ4C7MvM3sytV6k/bdwlUMEnP/ybw5cz8CPCXwBcj4iPAvcDJ\nzLwRONk8l7RHtIY/M89n5tPN4zeAl4AbgMPA8eZlx4E7Z1WkpP5d0Tl/RBwAPgr8BFjKzPNN06uM\nTgsk7RETX9sfEe8FHgG+lJm/jYg/tGVmRkTu8r5VYLVroZL6NVHPHxHvZhT8b2fmo83iCxGxr2nf\nB1zc6b2ZuZGZy5m53EfBkvrRGv4YdfEPAi9l5te2NZ0AjjaPjwKP91+epFmZ5LD/48BfA89FxDPN\nsvuAB4DvRcTdwC+Bu2ZToqRZaA1/Zv4XELs0f6LfciTNi1f4SUUZfqkowy8VZfilogy/VJThl4qK\nzB2vyp3Nxna5BFg1tU3Bvbm5OZ9CdrD98vW9JjMnKt6eXyrK8EtFGX6pKMMvFWX4paIMv1SU4ZeK\ncpxfC2ue/zcvt76+PrZ9bW1tPoVMwXF+SWMZfqkowy8VZfilogy/VJThl4oy/FJRjvNrz2q733/c\n9wVsbW2Nfe/KysoUFS0Gx/kljWX4paIMv1SU4ZeKMvxSUYZfKsrwS0W1jvNHxH7gIWAJSGAjM78e\nEWvA54FfNy+9LzN/0LIux/mlGZt0nH+S8O8D9mXm0xHxPuAp4E7gLuB3mfnPkxZl+KXZmzT875pg\nReeB883jNyLiJeCGbuVJGtoVnfNHxAHgo8BPmkX3RMSzEXEsIq7b5T2rEXE6Ik53qlRSrya+tj8i\n3gucAv4xMx+NiCXgNUafA/wDo1ODv21Zh4f90oz1ds4PEBHvBr4P/DAzv7ZD+wHg+5l5U8t6DL80\nY73d2BOj6UofBF7aHvzmg8BLPgs8f6VFShrOJJ/23wr8J/Ac8Haz+D7gCHAzo8P+M8AXmg8Hx63L\nnl+asV4P+/ti+KXZ835+SWMZfqkowy8VZfilogy/VJThl4oy/FJRhl8qyvBLRRl+qSjDLxVl+KWi\nDL9UlOGXimr9As+evQb8ctvz9zfLFtGi1raodYG1TavP2j446Qvnej//OzYecTozlwcrYIxFrW1R\n6wJrm9ZQtXnYLxVl+KWihg7/xsDbH2dRa1vUusDapjVIbYOe80saztA9v6SBDBL+iLg9In4aEa9E\nxL1D1LCbiDgTEc9FxDNDTzHWTIN2MSKe37bs+oj4cUS83PzecZq0gWpbi4hzzb57JiLuGKi2/RGx\nGREvRsQLEfF3zfJB992YugbZb3M/7I+Ia4CfAbcBZ4EngSOZ+eJcC9lFRJwBljNz8DHhiPgr4HfA\nQ5dmQ4qIfwJez8wHmj+c12Xm3y9IbWtc4czNM6ptt5ml/4YB912fM173YYie/xbglcz8RWb+HvgO\ncHiAOhZeZj4BvH7Z4sPA8ebxcUb/eeZul9oWQmaez8ynm8dvAJdmlh50342paxBDhP8G4Ffbnp9l\nsab8TuBHEfFURKwOXcwOlrbNjPQqsDRkMTtonbl5ni6bWXph9t00M173zQ/83unWzPwY8Gngi83h\n7ULK0TnbIg3XfAP4MKNp3M4DXx2ymGZm6UeAL2Xmb7e3DbnvdqhrkP02RPjPAfu3Pf9As2whZOa5\n5vdF4DFGpymL5MKlSVKb3xcHrucPMvNCZr6VmW8D32TAfdfMLP0I8O3MfLRZPPi+26muofbbEOF/\nErgxIj4UEe8BPgecGKCOd4iIa5sPYoiIa4FPsXizD58AjjaPjwKPD1jLH1mUmZt3m1magffdws14\nnZlz/wHuYPSJ/8+BrwxRwy51/TnwP83PC0PXBjzM6DDw/xh9NnI38KfASeBl4D+A6xeotn9lNJvz\ns4yCtm+g2m5ldEj/LPBM83PH0PtuTF2D7Dev8JOK8gM/qSjDLxVl+KWiDL9UlOGXijL8UlGGXyrK\n8EtF/T8tUQzyzj9HOwAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f56e03131d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Get one and predict\n",
    "r = random.randint(0, len(mnist_test) - 1)\n",
    "X_single_data = Variable(mnist_test.test_data[r:r + 1].view(-1, 28 * 28).float().div(255)).cuda()\n",
    "Y_single_data = Variable(mnist_test.test_labels[r:r + 1]).cuda()\n",
    "single_prediction = model(X_single_data)\n",
    "plt.imshow(X_single_data.cpu().data.view(28,28).numpy() , cmap='gray')\n",
    "\n",
    "print(\"Label: \", Y_single_data.data)\n",
    "print(\"Prediction: \", torch.max(single_prediction.data, 1)[1])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2017-08-07T06:40:06.006199Z",
     "start_time": "2017-08-07T06:41:37.130Z"
    }
   },
   "outputs": [
    {
     "ename": "TypeError",
     "evalue": "'RbfNet' object does not support indexing",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m--------------------------------------------------------\u001b[0m",
      "\u001b[0;31mTypeError\u001b[0m              Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-52-a319cb2a5318>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[1;32m      1\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mi\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mrange\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;36m20\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 2\u001b[0;31m     \u001b[0mweight\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mmodel\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mweight\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mi\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdata\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mview\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;36m28\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m28\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      3\u001b[0m     \u001b[0mweight\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0mweight\u001b[0m \u001b[0;34m-\u001b[0m \u001b[0mtorch\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmin\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mweight\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m/\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtorch\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmax\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mweight\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m-\u001b[0m\u001b[0mtorch\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmin\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mweight\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      4\u001b[0m     \u001b[0mplt\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mimshow\u001b[0m\u001b[0;34m(\u001b[0m \u001b[0mweight\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mnumpy\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mcmap\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m'gray'\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      5\u001b[0m     \u001b[0mplt\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mshow\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mTypeError\u001b[0m: 'RbfNet' object does not support indexing"
     ]
    }
   ],
   "source": [
    "for i in range(20):\n",
    "    weight = model[0].weight[i,:].data.view(28,28) \n",
    "    weight = (weight - torch.min(weight))/(torch.max(weight)-torch.min(weight))\n",
    "    plt.imshow( weight.numpy(), cmap='gray')\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python [default]",
   "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
}