Assignment 3.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 52,
   "outputs": [],
   "source": [
    "from __future__ import print_function\n",
    "import torch\n",
    "import torch.nn as nn\n",
    "import torch.optim as optim\n",
    "import pandas as pd\n",
    "import imgaug.augmenters as iaa\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "from mlxtend.data import loadlocal_mnist #to load data from local files\n"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n",
     "is_executing": false
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "outputs": [],
   "source": [
    "class MLP(nn.Module):\n",
    "    def __init__(self):\n",
    "        super(MLP, self).__init__()\n",
    "        self.fc1 = nn.Linear(28*28, 1500)\n",
    "        self.fc2 = nn.Linear(1500, 500)\n",
    "        self.fc22 = nn.Linear(500, 500)\n",
    "        self.fc3 = nn.Linear(500, 10)\n",
    "        self.relu = nn.ReLU()\n",
    "        self.log_softmax = nn.LogSoftmax(dim=1)\n",
    "\n",
    "    def forward(self, x):\n",
    "        x = self.relu(self.fc1(x))\n",
    "        x = self.relu(self.fc2(x))\n",
    "        x = self.relu(self.fc22(x))\n",
    "        x = self.fc3(x)\n",
    "        return self.log_softmax(x)\n",
    "    "
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n",
     "is_executing": false
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "outputs": [],
   "source": [
    "def train(log_interval, model, device, train_loader, optimizer, epoch):\n",
    "    model.train()\n",
    "    for batch_idx, (data, target) in enumerate(train_loader):\n",
    "        data, target = data.to(device), target.to(device)\n",
    "        optimizer.zero_grad()\n",
    "        output = model(data)\n",
    "        loss_f = nn.NLLLoss()\n",
    "        loss = loss_f(output, target)\n",
    "        loss.backward()\n",
    "        optimizer.step()\n",
    "        if batch_idx % log_interval == 0:\n",
    "            print('Train Epoch: {} [{}/{} ({:.0f}%)]\\tLoss: {:.6f}'.format(\n",
    "                epoch, batch_idx * len(data), len(train_loader.dataset),\n",
    "                100. * batch_idx / len(train_loader), loss.item()))"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n",
     "is_executing": false
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "outputs": [],
   "source": [
    "def test(model, device, test_loader):\n",
    "    model.eval()\n",
    "    test_loss = 0\n",
    "    correct = 0\n",
    "    loss_f = nn.NLLLoss(reduction='sum')\n",
    "    with torch.no_grad():\n",
    "        for batch_idx, (data, target) in enumerate(test_loader):\n",
    "            data, target = data.to(device), target.to(device)\n",
    "            output = model(data)\n",
    "            test_loss += loss_f(output, target).item() # sum up batch loss\n",
    "            pred = output.max(1, keepdim=True)[1] # get the index of the max log-probability\n",
    "            correct += pred.eq(target.view_as(pred)).sum().item()\n",
    "\n",
    "    test_loss /= len(test_loader.dataset)\n",
    "    return float(100. *correct / len(test_loader.dataset))"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n",
     "is_executing": false
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "outputs": [],
   "source": [
    "# Training settings\n",
    "batch_size = 1000\n",
    "test_batch_size = 1000\n",
    "epochs = 10\n",
    "lr = 0.001\n",
    "momentum = 0.9\n",
    "seed = 1\n",
    "log_interval = 800\n",
    "save_model = False\n",
    "use_cuda = True # change to False if running without GPU"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n",
     "is_executing": false
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "outputs": [],
   "source": [
    "torch.manual_seed(seed)\n",
    "\n",
    "device = torch.device(\"cuda\" if use_cuda else \"cpu\")"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n",
     "is_executing": false
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "outputs": [
    {
     "name": "stdout",
     "text": [
      "(60000, 784)\n",
      "data after augmentation: (240000, 28, 28)\n"
     ],
     "output_type": "stream"
    }
   ],
   "source": [
    "class TrainClass(torch.utils.data.Dataset):\n",
    "    def __init__(self, X, device):\n",
    "        self.data = torch.tensor(X[0], dtype=torch.float32).to(device)\n",
    "        # self.data.unsqueeze(0)\n",
    "        self.label = torch.tensor(X[1], dtype=torch.int64).to(device)\n",
    "        print(type(self.data))\n",
    "        print(self.label.type)\n",
    "    def __len__(self):\n",
    "        return len(self.data)\n",
    "    \n",
    "    def __getitem__(self, index):\n",
    "        data, target = self.data[index], self.label[index]\n",
    "        return data, target\n",
    "    \n",
    "kwargs = {'num_workers': 0, 'pin_memory': True} if use_cuda else {}\n",
    "\n",
    "X = loadlocal_mnist(\n",
    "    images_path='train-images-idx3-ubyte',\n",
    "    labels_path='train-labels-idx1-ubyte'\n",
    ")\n",
    "contrast = iaa.Sequential([\n",
    "    iaa.LinearContrast((0.75, 1.5)),\n",
    "    iaa.Crop(px=(0, 1))\n",
    "], random_order=True)\n",
    "affine = iaa.Sequential([\n",
    "    iaa.AdditiveGaussianNoise(loc=0, scale=(0.005*255, 0.05*255), per_channel=0.5),\n",
    "    iaa.Add((-5, 5)),\n",
    "    iaa.Affine(\n",
    "        scale={\"x\": (0.9, 1.05), \"y\": (0.9, 1.05)},\n",
    "        shear=(-1, 1)\n",
    "    )\n",
    "])\n",
    "X = list(X)\n",
    "def perform_augment(seq, data):\n",
    "    aug = seq(images=data[0])\n",
    "    data[0] = np.concatenate((data[0], aug), axis=0)\n",
    "    data[1] = np.concatenate((data[1], data[1]), axis=0)\n",
    "augments = [contrast, affine]\n",
    "print(X[0].shape)\n",
    "X[0] = X[0].reshape(X[0].shape[0], 28, 28)\n",
    "for x in augments:\n",
    "    perform_augment(x, X)\n",
    "print(\"data after augmentation: {}\".format(X[0].shape))"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n",
     "is_executing": false
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "outputs": [
    {
     "data": {
      "text/plain": "<Figure size 2016x2016 with 36 Axes>",
      "image/png": "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\n"
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#rotating etc didn't improve performance at all because images in test set are not rotated\n",
    "# example of how images looks like (4 created from one)\n",
    "n_rows = 9\n",
    "offset = 0\n",
    "images= X[0].reshape(X[0].shape[0], 28, 28)\n",
    "fig = plt.figure(figsize=(28, 28))\n",
    "columns = 4\n",
    "rows = n_rows\n",
    "for j in range(1, n_rows*columns+1):\n",
    "    img = images[offset+ int((j-1)/columns)+ int(j%columns)*60000]\n",
    "    fig.add_subplot(rows, columns, j)\n",
    "    plt.imshow(img, cmap='gray')\n",
    "plt.show()"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n",
     "is_executing": false
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "outputs": [
    {
     "name": "stdout",
     "text": [
      "<class 'torch.Tensor'>\n<built-in method type of Tensor object at 0x000002133F2DA1D8>\n"
     ],
     "output_type": "stream"
    }
   ],
   "source": [
    "X[0] = X[0].reshape(X[0].shape[0], 784)\n",
    "t = TrainClass(X, device='cpu')\n",
    "train_set, val_set = torch.utils.data.random_split(t, [int(X[0].shape[0]/6*5), int(X[0].shape[0]/6)])\n",
    "\n",
    "train_loader = torch.utils.data.DataLoader(\n",
    "    train_set,\n",
    "    batch_size = batch_size, shuffle=True, **kwargs)\n",
    "test_loader = torch.utils.data.DataLoader(\n",
    "    val_set,\n",
    "    batch_size = batch_size, shuffle=True, **kwargs)"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n",
     "is_executing": false
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "outputs": [],
   "source": [
    "model = MLP().to(device)\n",
    "optimizer = optim.Adam(model.parameters(), lr=lr)"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n",
     "is_executing": false
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "outputs": [
    {
     "name": "stdout",
     "text": [
      "Train Epoch: 1 [0/200000 (0%)]\tLoss: 5.443109\n",
      "Train Epoch: 2 [0/200000 (0%)]\tLoss: 0.433688\n",
      "Train Epoch: 3 [0/200000 (0%)]\tLoss: 0.345493\n",
      "Train Epoch: 4 [0/200000 (0%)]\tLoss: 0.285386\n",
      "Train Epoch: 5 [0/200000 (0%)]\tLoss: 0.318572\n",
      "Train Epoch: 6 [0/200000 (0%)]\tLoss: 0.268227\n",
      "Train Epoch: 7 [0/200000 (0%)]\tLoss: 0.268397\n",
      "Train Epoch: 8 [0/200000 (0%)]\tLoss: 0.202742\n",
      "Train Epoch: 9 [0/200000 (0%)]\tLoss: 0.234914\n",
      "Train Epoch: 10 [0/200000 (0%)]\tLoss: 0.230790\n"
     ],
     "output_type": "stream"
    }
   ],
   "source": [
    "test_accuracy = [0 for i in range(epochs+1)]\n",
    "#not trained model yet\n",
    "test_accuracy[0] = test(model, device, test_loader)\n",
    "for epoch in range(1, epochs + 1):\n",
    "    train(log_interval, model, device, train_loader, optimizer, epoch)\n",
    "    test_accuracy[epoch] = test(model, device, test_loader)"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n",
     "is_executing": false
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "outputs": [
    {
     "data": {
      "text/plain": "<Figure size 432x288 with 1 Axes>",
      "image/png": "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\n"
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(test_accuracy)\n",
    "plt.show()"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n",
     "is_executing": false
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "outputs": [
    {
     "name": "stdout",
     "text": [
      "Train Epoch: 11 [0/200000 (0%)]\tLoss: 0.221941\n",
      "Train Epoch: 12 [0/200000 (0%)]\tLoss: 0.219834\n",
      "Train Epoch: 13 [0/200000 (0%)]\tLoss: 0.220623\n",
      "Train Epoch: 14 [0/200000 (0%)]\tLoss: 0.207399\n",
      "Train Epoch: 15 [0/200000 (0%)]\tLoss: 0.190798\n",
      "Train Epoch: 16 [0/200000 (0%)]\tLoss: 0.180840\n",
      "Train Epoch: 17 [0/200000 (0%)]\tLoss: 0.183513\n",
      "Train Epoch: 18 [0/200000 (0%)]\tLoss: 0.176425\n",
      "Train Epoch: 19 [0/200000 (0%)]\tLoss: 0.188386\n",
      "Train Epoch: 20 [0/200000 (0%)]\tLoss: 0.163082\n",
      "Train Epoch: 21 [0/200000 (0%)]\tLoss: 0.218329\n",
      "Train Epoch: 22 [0/200000 (0%)]\tLoss: 0.141981\n",
      "Train Epoch: 23 [0/200000 (0%)]\tLoss: 0.196714\n",
      "Train Epoch: 24 [0/200000 (0%)]\tLoss: 0.156877\n",
      "Train Epoch: 25 [0/200000 (0%)]\tLoss: 0.128181\n",
      "Train Epoch: 26 [0/200000 (0%)]\tLoss: 0.126221\n",
      "Train Epoch: 27 [0/200000 (0%)]\tLoss: 0.186516\n",
      "Train Epoch: 28 [0/200000 (0%)]\tLoss: 0.146428\n",
      "Train Epoch: 29 [0/200000 (0%)]\tLoss: 0.140098\n",
      "Train Epoch: 30 [0/200000 (0%)]\tLoss: 0.154998\n"
     ],
     "output_type": "stream"
    },
    {
     "data": {
      "text/plain": "<Figure size 432x288 with 1 Axes>",
      "image/png": "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\n"
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#some more epochs\n",
    "test_accuracy = [0 for i in range(2*epochs)]\n",
    "for epoch in range(0, 2*epochs):\n",
    "    train(log_interval, model, device, train_loader, optimizer, epoch+11)\n",
    "    test_accuracy[epoch] = test(model, device, test_loader)"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n",
     "is_executing": false
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "outputs": [
    {
     "data": {
      "text/plain": "<Figure size 432x288 with 1 Axes>",
      "image/png": "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\n"
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#it is a bit too optimistic, because some of augmented images looks really similiar\n",
    "# and some of them might've ended up in training set, while others in test set.\n",
    "plt.plot(test_accuracy, label='test set result')\n",
    "plt.legend()\n",
    "plt.show()"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n",
     "is_executing": false
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "outputs": [
    {
     "name": "stdout",
     "text": [
      "Train Epoch: 1 [0/240000 (0%)]\tLoss: 5.439871\n",
      "Train Epoch: 2 [0/240000 (0%)]\tLoss: 0.406597\n",
      "Train Epoch: 3 [0/240000 (0%)]\tLoss: 0.327881\n",
      "Train Epoch: 4 [0/240000 (0%)]\tLoss: 0.315861\n",
      "Train Epoch: 5 [0/240000 (0%)]\tLoss: 0.292652\n",
      "Train Epoch: 6 [0/240000 (0%)]\tLoss: 0.253505\n",
      "Train Epoch: 7 [0/240000 (0%)]\tLoss: 0.240194\n",
      "Train Epoch: 8 [0/240000 (0%)]\tLoss: 0.257415\n",
      "Train Epoch: 9 [0/240000 (0%)]\tLoss: 0.240968\n",
      "Train Epoch: 10 [0/240000 (0%)]\tLoss: 0.229273\n",
      "Train Epoch: 11 [0/240000 (0%)]\tLoss: 0.200161\n",
      "Train Epoch: 12 [0/240000 (0%)]\tLoss: 0.212372\n",
      "Train Epoch: 13 [0/240000 (0%)]\tLoss: 0.171307\n",
      "Train Epoch: 14 [0/240000 (0%)]\tLoss: 0.204088\n",
      "Train Epoch: 15 [0/240000 (0%)]\tLoss: 0.214935\n",
      "Train Epoch: 16 [0/240000 (0%)]\tLoss: 0.169584\n",
      "Train Epoch: 17 [0/240000 (0%)]\tLoss: 0.193845\n",
      "Train Epoch: 18 [0/240000 (0%)]\tLoss: 0.164667\n",
      "Train Epoch: 19 [0/240000 (0%)]\tLoss: 0.147247\n",
      "Train Epoch: 20 [0/240000 (0%)]\tLoss: 0.166615\n",
      "Train Epoch: 21 [0/240000 (0%)]\tLoss: 0.186231\n",
      "Train Epoch: 22 [0/240000 (0%)]\tLoss: 0.136774\n",
      "Train Epoch: 23 [0/240000 (0%)]\tLoss: 0.149698\n",
      "Train Epoch: 24 [0/240000 (0%)]\tLoss: 0.134606\n",
      "Train Epoch: 25 [0/240000 (0%)]\tLoss: 0.145346\n",
      "Train Epoch: 26 [0/240000 (0%)]\tLoss: 0.169041\n",
      "Train Epoch: 27 [0/240000 (0%)]\tLoss: 0.144492\n",
      "Train Epoch: 28 [0/240000 (0%)]\tLoss: 0.141756\n",
      "Train Epoch: 29 [0/240000 (0%)]\tLoss: 0.126924\n",
      "Train Epoch: 30 [0/240000 (0%)]\tLoss: 0.121306\n",
      "Train Epoch: 31 [0/240000 (0%)]\tLoss: 0.129334\n",
      "Train Epoch: 32 [0/240000 (0%)]\tLoss: 0.129325\n",
      "Train Epoch: 33 [0/240000 (0%)]\tLoss: 0.113790\n",
      "Train Epoch: 34 [0/240000 (0%)]\tLoss: 0.145132\n",
      "Train Epoch: 35 [0/240000 (0%)]\tLoss: 0.126871\n",
      "Train Epoch: 36 [0/240000 (0%)]\tLoss: 0.107384\n",
      "Train Epoch: 37 [0/240000 (0%)]\tLoss: 0.102979\n",
      "Train Epoch: 38 [0/240000 (0%)]\tLoss: 0.099719\n",
      "Train Epoch: 39 [0/240000 (0%)]\tLoss: 0.120410\n",
      "Train Epoch: 40 [0/240000 (0%)]\tLoss: 0.113280\n",
      "Train Epoch: 41 [0/240000 (0%)]\tLoss: 0.085732\n",
      "Train Epoch: 42 [0/240000 (0%)]\tLoss: 0.131488\n",
      "Train Epoch: 43 [0/240000 (0%)]\tLoss: 0.120756\n",
      "Train Epoch: 44 [0/240000 (0%)]\tLoss: 0.080988\n",
      "Train Epoch: 45 [0/240000 (0%)]\tLoss: 0.118599\n",
      "Train Epoch: 46 [0/240000 (0%)]\tLoss: 0.102519\n",
      "Train Epoch: 47 [0/240000 (0%)]\tLoss: 0.075327\n",
      "Train Epoch: 48 [0/240000 (0%)]\tLoss: 0.117993\n",
      "Train Epoch: 49 [0/240000 (0%)]\tLoss: 0.089408\n",
      "Train Epoch: 50 [0/240000 (0%)]\tLoss: 0.110016\n",
      "Train Epoch: 51 [0/240000 (0%)]\tLoss: 0.077494\n",
      "Train Epoch: 52 [0/240000 (0%)]\tLoss: 0.105720\n",
      "Train Epoch: 53 [0/240000 (0%)]\tLoss: 0.062903\n",
      "Train Epoch: 54 [0/240000 (0%)]\tLoss: 0.086381\n",
      "Train Epoch: 55 [0/240000 (0%)]\tLoss: 0.068912\n",
      "Train Epoch: 56 [0/240000 (0%)]\tLoss: 0.075359\n",
      "Train Epoch: 57 [0/240000 (0%)]\tLoss: 0.081313\n",
      "Train Epoch: 58 [0/240000 (0%)]\tLoss: 0.063524\n",
      "Train Epoch: 59 [0/240000 (0%)]\tLoss: 0.078540\n",
      "Train Epoch: 60 [0/240000 (0%)]\tLoss: 0.061121\n",
      "Train Epoch: 61 [0/240000 (0%)]\tLoss: 0.071914\n",
      "Train Epoch: 62 [0/240000 (0%)]\tLoss: 0.099853\n",
      "Train Epoch: 63 [0/240000 (0%)]\tLoss: 0.080538\n",
      "Train Epoch: 64 [0/240000 (0%)]\tLoss: 0.083717\n",
      "Train Epoch: 65 [0/240000 (0%)]\tLoss: 0.070466\n",
      "Train Epoch: 66 [0/240000 (0%)]\tLoss: 0.090929\n",
      "Train Epoch: 67 [0/240000 (0%)]\tLoss: 0.058396\n",
      "Train Epoch: 68 [0/240000 (0%)]\tLoss: 0.066675\n",
      "Train Epoch: 69 [0/240000 (0%)]\tLoss: 0.102712\n",
      "Train Epoch: 70 [0/240000 (0%)]\tLoss: 0.103492\n",
      "Train Epoch: 71 [0/240000 (0%)]\tLoss: 0.063295\n",
      "Train Epoch: 72 [0/240000 (0%)]\tLoss: 0.065611\n",
      "Train Epoch: 73 [0/240000 (0%)]\tLoss: 0.069915\n",
      "Train Epoch: 74 [0/240000 (0%)]\tLoss: 0.071641\n",
      "Train Epoch: 75 [0/240000 (0%)]\tLoss: 0.083776\n",
      "Train Epoch: 76 [0/240000 (0%)]\tLoss: 0.056702\n",
      "Train Epoch: 77 [0/240000 (0%)]\tLoss: 0.090994\n",
      "Train Epoch: 78 [0/240000 (0%)]\tLoss: 0.077501\n",
      "Train Epoch: 79 [0/240000 (0%)]\tLoss: 0.088724\n",
      "Train Epoch: 80 [0/240000 (0%)]\tLoss: 0.096543\n",
      "Train Epoch: 81 [0/240000 (0%)]\tLoss: 0.079194\n",
      "Train Epoch: 82 [0/240000 (0%)]\tLoss: 0.039740\n",
      "Train Epoch: 83 [0/240000 (0%)]\tLoss: 0.080677\n",
      "Train Epoch: 84 [0/240000 (0%)]\tLoss: 0.086982\n",
      "Train Epoch: 85 [0/240000 (0%)]\tLoss: 0.051938\n",
      "Train Epoch: 86 [0/240000 (0%)]\tLoss: 0.080751\n",
      "Train Epoch: 87 [0/240000 (0%)]\tLoss: 0.061627\n",
      "Train Epoch: 88 [0/240000 (0%)]\tLoss: 0.049092\n",
      "Train Epoch: 89 [0/240000 (0%)]\tLoss: 0.073519\n",
      "Train Epoch: 90 [0/240000 (0%)]\tLoss: 0.052620\n",
      "Train Epoch: 91 [0/240000 (0%)]\tLoss: 0.055288\n",
      "Train Epoch: 92 [0/240000 (0%)]\tLoss: 0.088253\n",
      "Train Epoch: 93 [0/240000 (0%)]\tLoss: 0.076353\n",
      "Train Epoch: 94 [0/240000 (0%)]\tLoss: 0.057072\n",
      "Train Epoch: 95 [0/240000 (0%)]\tLoss: 0.055498\n",
      "Train Epoch: 96 [0/240000 (0%)]\tLoss: 0.054900\n",
      "Train Epoch: 97 [0/240000 (0%)]\tLoss: 0.049901\n",
      "Train Epoch: 98 [0/240000 (0%)]\tLoss: 0.080622\n",
      "Train Epoch: 99 [0/240000 (0%)]\tLoss: 0.040132\n",
      "Train Epoch: 100 [0/240000 (0%)]\tLoss: 0.051425\n"
     ],
     "output_type": "stream"
    }
   ],
   "source": [
    "model = MLP().to(device)\n",
    "optimizer = optim.Adam(model.parameters(), lr=lr)\n",
    "#train on whole set\n",
    "train_whole_set_loader = torch.utils.data.DataLoader(\n",
    "    t,\n",
    "    batch_size = batch_size, shuffle=True, **kwargs\n",
    ")\n",
    "max_epochs = 100\n",
    "train_accuracy = [0 for i in range(max_epochs)]\n",
    "for epoch in range(1, max_epochs + 1):\n",
    "    train(log_interval, model, device, train_whole_set_loader, optimizer, epoch)\n",
    "    train_accuracy[epoch-1] = test(model, device, test_loader) "
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n",
     "is_executing": false
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "outputs": [
    {
     "data": {
      "text/plain": "<Figure size 432x288 with 1 Axes>",
      "image/png": "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\n"
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(train_accuracy, label='train set result')\n",
    "plt.legend()\n",
    "plt.show()"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n",
     "is_executing": false
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "outputs": [
    {
     "name": "stdout",
     "text": [
      "<class 'torch.Tensor'>\n<built-in method type of Tensor object at 0x000002133A5749A8>\n"
     ],
     "output_type": "stream"
    }
   ],
   "source": [
    "#needed to create file with some random labels to load test data,\n",
    "# because loadlocal requires you to have labels file for some reason.\n",
    "t_test = loadlocal_mnist('t10k-images-idx3-ubyte', 't10k-labels.idx1-ubyte')\n",
    "t_test = list(t_test)\n",
    "tt = TrainClass(t_test, device=device)\n",
    "\n",
    "outputs = model(tt.data)\n",
    "outputs = outputs.argmax(dim=1)\n",
    "result = pd.DataFrame()\n",
    "result['Id'] = [i for i in range (0, 10000)]\n",
    "result['Class'] = outputs.cpu()\n",
    "result.to_csv(\"nn_wynik.csv\", index=False, header=True)"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n",
     "is_executing": false
    }
   }
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.6"
  },
  "pycharm": {
   "stem_cell": {
    "cell_type": "raw",
    "source": [],
    "metadata": {
     "collapsed": false
    }
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}