L5 Part 5 - Inference and Validation.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Inference and Validation"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "__Inference__ : making predictions using the trained network(model)\n",
    "\n",
    "__Validation__ : Checking the performance of model on data not seen before\n",
    "\n",
    "__Aim__ : is to minimise the Validation loss (not the training loss)\n",
    "\n",
    "Only minimising the training loss can result in overfitting.\n",
    "\n",
    "So while training, we keep a check on the performance of the network on data not in the training set (called Validation set), to __check overfitting__ (__Validation__)\n",
    "\n",
    "Ways to __reduce overfitting__ :\n",
    "1. __Early Stopping__ : stopping on a threshold of epochs - so that overfitting does not occur\n",
    "2. __Dropout__ : randomly dropping out input features - so that the model does not overfit the training data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Loading the Fashion-MNIST data using torchvision library\n",
    "\n",
    "import torch\n",
    "from torchvision import datasets, transforms\n",
    "\n",
    "# Define a transform to normalize the data\n",
    "transform = transforms.Compose([transforms.ToTensor(),\n",
    "                                transforms.Normalize((0.5,), (0.5,))])\n",
    "# Download and load the training data\n",
    "trainset = datasets.FashionMNIST('~/.pytorch/F_MNIST_data/', download=True, train=True, transform=transform)\n",
    "trainloader = torch.utils.data.DataLoader(trainset, batch_size=64, shuffle=True)\n",
    "\n",
    "# Download and load the test data\n",
    "testset = datasets.FashionMNIST('~/.pytorch/F_MNIST_data/', download=True, train=False, transform=transform)\n",
    "testloader = torch.utils.data.DataLoader(testset, batch_size=64, shuffle=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "torch.Size([64, 1, 28, 28])\n",
      "torch.Size([64])\n"
     ]
    }
   ],
   "source": [
    "# Checking the shape of data\n",
    "\n",
    "images,labels = next(iter(trainloader))\n",
    "\n",
    "print(images.shape)\n",
    "print(labels.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Defining a model to train on this data\n",
    "\n",
    "from torch import nn, optim\n",
    "import torch.nn.functional as F\n",
    "\n",
    "class Classifier(nn.Module):\n",
    "    \n",
    "    def __init__(self):\n",
    "        \n",
    "        super().__init__()\n",
    "        \n",
    "        # 3 hidden layers, 1 output layer\n",
    "        self.fc1 = nn.Linear(784,256)\n",
    "        self.fc2 = nn.Linear(256,128)\n",
    "        self.fc3 = nn.Linear(128,64)\n",
    "        self.fc4 = nn.Linear(64,10)\n",
    "        \n",
    "    def forward(self, x):\n",
    "        \n",
    "        # must flatten the input tensor\n",
    "        x = x.view(x.shape[0], -1)\n",
    "        \n",
    "        x = F.relu(self.fc1(x))\n",
    "        x = F.relu(self.fc2(x))\n",
    "        x = F.relu(self.fc3(x))\n",
    "        x = F.log_softmax(self.fc4(x), dim=1)\n",
    "        \n",
    "        return x"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch:  1   Training loss:  0.50949853251992   Test loss:  tensor(0.4427)   Test Accuracy:  tensor(0.8379)\n",
      "Epoch:  2   Training loss:  0.3911836894908185   Test loss:  tensor(0.4563)   Test Accuracy:  tensor(0.8398)\n",
      "Epoch:  3   Training loss:  0.35544839537124645   Test loss:  tensor(0.4339)   Test Accuracy:  tensor(0.8418)\n",
      "Epoch:  4   Training loss:  0.3333609958431487   Test loss:  tensor(0.3718)   Test Accuracy:  tensor(0.8701)\n",
      "Epoch:  5   Training loss:  0.31556908676285605   Test loss:  tensor(0.3757)   Test Accuracy:  tensor(0.8648)\n",
      "Epoch:  6   Training loss:  0.30435090843262447   Test loss:  tensor(0.3598)   Test Accuracy:  tensor(0.8721)\n",
      "Epoch:  7   Training loss:  0.29517025422694076   Test loss:  tensor(0.3770)   Test Accuracy:  tensor(0.8710)\n",
      "Epoch:  8   Training loss:  0.2815995504543471   Test loss:  tensor(0.3917)   Test Accuracy:  tensor(0.8650)\n",
      "Epoch:  9   Training loss:  0.2725007716915818   Test loss:  tensor(0.3668)   Test Accuracy:  tensor(0.8721)\n",
      "Epoch:  10   Training loss:  0.27107518388709023   Test loss:  tensor(0.3702)   Test Accuracy:  tensor(0.8766)\n",
      "Epoch:  11   Training loss:  0.25775898270197767   Test loss:  tensor(0.3798)   Test Accuracy:  tensor(0.8717)\n",
      "Epoch:  12   Training loss:  0.25405623556486073   Test loss:  tensor(0.3818)   Test Accuracy:  tensor(0.8783)\n",
      "Epoch:  13   Training loss:  0.24637530029201304   Test loss:  tensor(0.3740)   Test Accuracy:  tensor(0.8769)\n",
      "Epoch:  14   Training loss:  0.24309211028521377   Test loss:  tensor(0.3809)   Test Accuracy:  tensor(0.8759)\n",
      "Epoch:  15   Training loss:  0.24192441002265222   Test loss:  tensor(0.3718)   Test Accuracy:  tensor(0.8776)\n",
      "Epoch:  16   Training loss:  0.23745231609195788   Test loss:  tensor(0.3639)   Test Accuracy:  tensor(0.8855)\n",
      "Epoch:  17   Training loss:  0.22590806780815886   Test loss:  tensor(0.3676)   Test Accuracy:  tensor(0.8835)\n",
      "Epoch:  18   Training loss:  0.22519999120169992   Test loss:  tensor(0.4169)   Test Accuracy:  tensor(0.8778)\n",
      "Epoch:  19   Training loss:  0.22512833318158762   Test loss:  tensor(0.4079)   Test Accuracy:  tensor(0.8769)\n",
      "Epoch:  20   Training loss:  0.2161990506117786   Test loss:  tensor(0.3728)   Test Accuracy:  tensor(0.8748)\n",
      "Epoch:  21   Training loss:  0.21341899115600185   Test loss:  tensor(0.3859)   Test Accuracy:  tensor(0.8778)\n",
      "Epoch:  22   Training loss:  0.21476702777339196   Test loss:  tensor(0.3972)   Test Accuracy:  tensor(0.8743)\n",
      "Epoch:  23   Training loss:  0.2070506290395631   Test loss:  tensor(0.3841)   Test Accuracy:  tensor(0.8851)\n",
      "Epoch:  24   Training loss:  0.20201751111603494   Test loss:  tensor(0.4127)   Test Accuracy:  tensor(0.8795)\n",
      "Epoch:  25   Training loss:  0.20500864026976673   Test loss:  tensor(0.3995)   Test Accuracy:  tensor(0.8841)\n",
      "Epoch:  26   Training loss:  0.20752415217077935   Test loss:  tensor(0.4048)   Test Accuracy:  tensor(0.8761)\n",
      "Epoch:  27   Training loss:  0.19584699293602503   Test loss:  tensor(0.3981)   Test Accuracy:  tensor(0.8827)\n",
      "Epoch:  28   Training loss:  0.1865891503201905   Test loss:  tensor(0.3964)   Test Accuracy:  tensor(0.8828)\n",
      "Epoch:  29   Training loss:  0.19195797432190193   Test loss:  tensor(0.4153)   Test Accuracy:  tensor(0.8820)\n",
      "Epoch:  30   Training loss:  0.1855959869992695   Test loss:  tensor(0.4191)   Test Accuracy:  tensor(0.8764)\n"
     ]
    },
    {
     "ename": "KeyboardInterrupt",
     "evalue": "",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m                         Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-19-a543d33f57eb>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m     18\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     19\u001b[0m     \u001b[0;31m# Training on Training set\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 20\u001b[0;31m     \u001b[0;32mfor\u001b[0m \u001b[0mimages\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mlabels\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mtrainloader\u001b[0m\u001b[0;34m:\u001b[0m  \u001b[0;31m# (images, labels): batch of 64\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     21\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     22\u001b[0m         \u001b[0mlogps\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mmodel\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mforward\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mimages\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/anaconda3/lib/python3.6/site-packages/torch/utils/data/dataloader.py\u001b[0m in \u001b[0;36m__next__\u001b[0;34m(self)\u001b[0m\n\u001b[1;32m    344\u001b[0m     \u001b[0;32mdef\u001b[0m \u001b[0m__next__\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    345\u001b[0m         \u001b[0mindex\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_next_index\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m  \u001b[0;31m# may raise StopIteration\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 346\u001b[0;31m         \u001b[0mdata\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_dataset_fetcher\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfetch\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mindex\u001b[0m\u001b[0;34m)\u001b[0m  \u001b[0;31m# may raise StopIteration\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    347\u001b[0m         \u001b[0;32mif\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_pin_memory\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    348\u001b[0m             \u001b[0mdata\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0m_utils\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mpin_memory\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mpin_memory\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdata\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/anaconda3/lib/python3.6/site-packages/torch/utils/data/_utils/fetch.py\u001b[0m in \u001b[0;36mfetch\u001b[0;34m(self, possibly_batched_index)\u001b[0m\n\u001b[1;32m     42\u001b[0m     \u001b[0;32mdef\u001b[0m \u001b[0mfetch\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mpossibly_batched_index\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     43\u001b[0m         \u001b[0;32mif\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mauto_collation\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 44\u001b[0;31m             \u001b[0mdata\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdataset\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0midx\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0midx\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mpossibly_batched_index\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     45\u001b[0m         \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     46\u001b[0m             \u001b[0mdata\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdataset\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mpossibly_batched_index\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/anaconda3/lib/python3.6/site-packages/torch/utils/data/_utils/fetch.py\u001b[0m in \u001b[0;36m<listcomp>\u001b[0;34m(.0)\u001b[0m\n\u001b[1;32m     42\u001b[0m     \u001b[0;32mdef\u001b[0m \u001b[0mfetch\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mpossibly_batched_index\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     43\u001b[0m         \u001b[0;32mif\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mauto_collation\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 44\u001b[0;31m             \u001b[0mdata\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdataset\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0midx\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0midx\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mpossibly_batched_index\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     45\u001b[0m         \u001b[0;32melse\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     46\u001b[0m             \u001b[0mdata\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdataset\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mpossibly_batched_index\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m~/anaconda3/lib/python3.6/site-packages/torchvision/datasets/mnist.py\u001b[0m in \u001b[0;36m__getitem__\u001b[0;34m(self, index)\u001b[0m\n\u001b[1;32m     95\u001b[0m             \u001b[0mimg\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtransform\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mimg\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     96\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 97\u001b[0;31m         \u001b[0;32mif\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtarget_transform\u001b[0m \u001b[0;32mis\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     98\u001b[0m             \u001b[0mtarget\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtarget_transform\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtarget\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     99\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mKeyboardInterrupt\u001b[0m: "
     ]
    }
   ],
   "source": [
    "# Creating the model and training it\n",
    "\n",
    "model = Classifier()\n",
    "\n",
    "criterion = nn.NLLLoss()\n",
    "optimizer = optim.Adam(model.parameters(), lr = 0.003) # Similar to SGD optimizer, but is better\n",
    "\n",
    "epochs = 30\n",
    "\n",
    "train_losses = []\n",
    "test_losses = []\n",
    "\n",
    "for e in range(epochs):\n",
    "    \n",
    "    training_loss_per_epoch = 0\n",
    "    test_loss_per_epoch = 0\n",
    "    accuracy = 0\n",
    "    \n",
    "    # Training on Training set\n",
    "    for images, labels in trainloader:  # (images, labels): batch of 64\n",
    "        \n",
    "        logps = model.forward(images)\n",
    "        loss = criterion(logps,labels)\n",
    "        \n",
    "        optimizer.zero_grad()\n",
    "        loss.backward()\n",
    "        optimizer.step()\n",
    "        \n",
    "        training_loss_per_epoch += loss.item()\n",
    "        \n",
    "    else:\n",
    "        \n",
    "        # Validation using Test set\n",
    "        # turn off gradients while validation (since we are not training), increases speed and decreases memory \n",
    "        with torch.no_grad():\n",
    "            \n",
    "            for images, labels in testloader:\n",
    "\n",
    "                logps = model(images)\n",
    "                test_loss_per_epoch += criterion(logps, labels)\n",
    "                \n",
    "                # cummulative accuracy over all batches of the test set\n",
    "                ps = torch.exp(logps)\n",
    "                top_p, top_class = ps.topk(k=1, dim=1)\n",
    "                equals = top_class == labels.view(top_class.shape)\n",
    "                accuracy += torch.mean(equals.type(torch.FloatTensor))\n",
    "              \n",
    "        train_losses.append(training_loss_per_epoch/len(trainloader))\n",
    "        test_losses.append(test_loss_per_epoch/len(testloader))\n",
    "        \n",
    "        print(\"Epoch: \", e+1, \"  Training loss: \", training_loss_per_epoch/len(trainloader), \n",
    "              \"  Test loss: \", test_loss_per_epoch/len(testloader), \n",
    "              \"  Test Accuracy: \", accuracy/len(testloader) )\n",
    "        \n",
    "        "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x7fa9ebc4b6d8>"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "image/png": {
       "height": 248,
       "width": 380
      },
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Plotting how the training and validation (test) losses vary while training the model/network\n",
    "\n",
    "%matplotlib inline\n",
    "%config InlineBackend.figure_format = 'retina'\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "plt.plot(train_losses, label='Training loss')\n",
    "plt.plot(test_losses, label='Validation loss')\n",
    "plt.legend(frameon=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Using __Dropout__ : to reduce Overfitting"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Defining a new model to train : using dropout\n",
    "\n",
    "from torch import nn, optim\n",
    "import torch.nn.functional as F\n",
    "\n",
    "class Classifier(nn.Module):\n",
    "    \n",
    "    def __init__(self):\n",
    "        \n",
    "        super().__init__()\n",
    "        \n",
    "        # 3 hidden layers, 1 output layer\n",
    "        self.fc1 = nn.Linear(784,256)\n",
    "        self.fc2 = nn.Linear(256,128)\n",
    "        self.fc3 = nn.Linear(128,64)\n",
    "        self.fc4 = nn.Linear(64,10)\n",
    "        \n",
    "        # to reduce overfitting\n",
    "        self.dropout = nn.Dropout(p=0.2)\n",
    "        \n",
    "    def forward(self, x):\n",
    "        \n",
    "        # must flatten the input tensor\n",
    "        x = x.view(x.shape[0], -1)\n",
    "        \n",
    "        x = self.dropout(F.relu(self.fc1(x)))\n",
    "        x = self.dropout(F.relu(self.fc2(x)))\n",
    "        x = self.dropout(F.relu(self.fc3(x)))\n",
    "        x = F.log_softmax(self.fc4(x), dim=1)\n",
    "        \n",
    "        return x"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch:  1   Training loss:  0.6034254679547698   Test loss:  tensor(0.5319)   Test Accuracy:  tensor(0.8194)\n",
      "Epoch:  2   Training loss:  0.47972812851481855   Test loss:  tensor(0.4885)   Test Accuracy:  tensor(0.8337)\n",
      "Epoch:  3   Training loss:  0.4520124775101381   Test loss:  tensor(0.4807)   Test Accuracy:  tensor(0.8309)\n",
      "Epoch:  4   Training loss:  0.4343854394644054   Test loss:  tensor(0.4806)   Test Accuracy:  tensor(0.8343)\n",
      "Epoch:  5   Training loss:  0.4198857650859778   Test loss:  tensor(0.4704)   Test Accuracy:  tensor(0.8352)\n",
      "Epoch:  6   Training loss:  0.41095273220526385   Test loss:  tensor(0.4759)   Test Accuracy:  tensor(0.8438)\n",
      "Epoch:  7   Training loss:  0.4033419558011901   Test loss:  tensor(0.4858)   Test Accuracy:  tensor(0.8477)\n",
      "Epoch:  8   Training loss:  0.3976504465283107   Test loss:  tensor(0.4559)   Test Accuracy:  tensor(0.8434)\n",
      "Epoch:  9   Training loss:  0.3925067287391183   Test loss:  tensor(0.4858)   Test Accuracy:  tensor(0.8387)\n",
      "Epoch:  10   Training loss:  0.39066149213357265   Test loss:  tensor(0.4503)   Test Accuracy:  tensor(0.8464)\n",
      "Epoch:  11   Training loss:  0.3850566005624179   Test loss:  tensor(0.4721)   Test Accuracy:  tensor(0.8431)\n",
      "Epoch:  12   Training loss:  0.37902121922608883   Test loss:  tensor(0.5151)   Test Accuracy:  tensor(0.8293)\n",
      "Epoch:  13   Training loss:  0.3774590903460217   Test loss:  tensor(0.4511)   Test Accuracy:  tensor(0.8495)\n",
      "Epoch:  14   Training loss:  0.3745600552296143   Test loss:  tensor(0.4510)   Test Accuracy:  tensor(0.8550)\n",
      "Epoch:  15   Training loss:  0.3676422257174942   Test loss:  tensor(0.4633)   Test Accuracy:  tensor(0.8514)\n",
      "Epoch:  16   Training loss:  0.3852904706176664   Test loss:  tensor(0.4583)   Test Accuracy:  tensor(0.8447)\n",
      "Epoch:  17   Training loss:  0.36285366545290326   Test loss:  tensor(0.4532)   Test Accuracy:  tensor(0.8549)\n",
      "Epoch:  18   Training loss:  0.3643802019086347   Test loss:  tensor(0.4408)   Test Accuracy:  tensor(0.8524)\n",
      "Epoch:  19   Training loss:  0.3606131796611906   Test loss:  tensor(0.4576)   Test Accuracy:  tensor(0.8469)\n",
      "Epoch:  20   Training loss:  0.3579006899775727   Test loss:  tensor(0.4791)   Test Accuracy:  tensor(0.8533)\n",
      "Epoch:  21   Training loss:  0.357603976332239   Test loss:  tensor(0.4828)   Test Accuracy:  tensor(0.8500)\n",
      "Epoch:  22   Training loss:  0.35896545816173175   Test loss:  tensor(0.4585)   Test Accuracy:  tensor(0.8491)\n",
      "Epoch:  23   Training loss:  0.35331081063635567   Test loss:  tensor(0.4483)   Test Accuracy:  tensor(0.8537)\n",
      "Epoch:  24   Training loss:  0.35619942398308946   Test loss:  tensor(0.4418)   Test Accuracy:  tensor(0.8563)\n",
      "Epoch:  25   Training loss:  0.3460487978560711   Test loss:  tensor(0.4605)   Test Accuracy:  tensor(0.8525)\n",
      "Epoch:  26   Training loss:  0.3487608299485402   Test loss:  tensor(0.4502)   Test Accuracy:  tensor(0.8529)\n",
      "Epoch:  27   Training loss:  0.34454529675275786   Test loss:  tensor(0.4661)   Test Accuracy:  tensor(0.8606)\n",
      "Epoch:  28   Training loss:  0.3439900754913211   Test loss:  tensor(0.4647)   Test Accuracy:  tensor(0.8562)\n",
      "Epoch:  29   Training loss:  0.3477600347766998   Test loss:  tensor(0.4634)   Test Accuracy:  tensor(0.8523)\n",
      "Epoch:  30   Training loss:  0.3427540381540304   Test loss:  tensor(0.4441)   Test Accuracy:  tensor(0.8597)\n"
     ]
    }
   ],
   "source": [
    "# Training this new model on the train data, and simultaneously performing validation after every epoch\n",
    "\n",
    "model = Classifier()\n",
    "\n",
    "criterion = nn.NLLLoss()\n",
    "optimizer = optim.Adam(model.parameters(), lr = 0.003)\n",
    "\n",
    "epochs = 30\n",
    "\n",
    "train_losses = []\n",
    "test_losses = []\n",
    "\n",
    "for e in range(epochs):\n",
    "    \n",
    "    training_loss_per_epoch = 0\n",
    "    test_loss_per_epoch = 0\n",
    "    accuracy = 0\n",
    "    \n",
    "    # Training on Training set\n",
    "    for images, labels in trainloader:  # (images, labels): batch of 64\n",
    "        \n",
    "        logps = model.forward(images)\n",
    "        loss = criterion(logps,labels)\n",
    "        \n",
    "        optimizer.zero_grad()\n",
    "        loss.backward()\n",
    "        optimizer.step()\n",
    "        \n",
    "        training_loss_per_epoch += loss.item()\n",
    "        \n",
    "    else:\n",
    "        \n",
    "        # Validation using Test set\n",
    "        \n",
    "        # turn off gradients while validation (since we are not training), increases speed and decreases memory \n",
    "        with torch.no_grad():\n",
    "            \n",
    "            model.eval()  # turns off dropout\n",
    "            \n",
    "            for images, labels in testloader:\n",
    "\n",
    "                logps = model(images)\n",
    "                test_loss_per_epoch += criterion(logps, labels)\n",
    "                \n",
    "                # cummulative accuracy over all batches of the test set\n",
    "                ps = torch.exp(logps)\n",
    "                top_p, top_class = ps.topk(k=1, dim=1)\n",
    "                equals = top_class == labels.view(top_class.shape)\n",
    "                accuracy += torch.mean(equals.type(torch.FloatTensor))\n",
    "              \n",
    "        train_losses.append(training_loss_per_epoch/len(trainloader))\n",
    "        test_losses.append(test_loss_per_epoch/len(testloader))\n",
    "        \n",
    "        print(\"Epoch: \", e+1, \"  Training loss: \", training_loss_per_epoch/len(trainloader), \n",
    "              \"  Test loss: \", test_loss_per_epoch/len(testloader), \n",
    "              \"  Test Accuracy: \", accuracy/len(testloader) )\n",
    "        "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x7fa9ebc2a780>"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "image/png": {
       "height": 248,
       "width": 380
      },
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Plot : Training loss vs Validation/Test Loss (when using dropout technique)\n",
    "\n",
    "%matplotlib inline\n",
    "%config InlineBackend.figure_format = 'retina'\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "plt.plot(train_losses, label='Training loss')\n",
    "plt.plot(test_losses, label='Validation loss')\n",
    "plt.legend(frameon=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Better. Overfitting reduced - both training and validation loss decreasing!\n",
    "\n",
    "Not as in the model without dropout - where the training loss kept decreasing, but the validation loss was increasing."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Inference\n",
    "\n",
    "Now that the model is trained, we can use it for inference. We've done this before, but now we need to remember to set the model in inference mode with `model.eval()` - __to turn off dropout__. You'll also want to __turn off autograd__ with the `torch.no_grad()` context."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "tensor(6)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "Classifier(\n",
       "  (fc1): Linear(in_features=784, out_features=256, bias=True)\n",
       "  (fc2): Linear(in_features=256, out_features=128, bias=True)\n",
       "  (fc3): Linear(in_features=128, out_features=64, bias=True)\n",
       "  (fc4): Linear(in_features=64, out_features=10, bias=True)\n",
       "  (dropout): Dropout(p=0.2, inplace=False)\n",
       ")"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x648 with 2 Axes>"
      ]
     },
     "metadata": {
      "image/png": {
       "height": 216,
       "width": 424
      },
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Setting the trained model to evaluation mode - to turn off dropout - since we are not training\n",
    "model.eval() # turn off dropouts\n",
    "\n",
    "# grabbing a batch of test images\n",
    "images, labels = next(iter(testloader))\n",
    "\n",
    "# Picking one image\n",
    "img = images[0]\n",
    "\n",
    "# Using the trained model/network (flattening of input - done within the model)\n",
    "\n",
    "with torch.no_grad(): # turnn off the autograd - since no training is done here\n",
    "    log_ps = model(img)\n",
    "\n",
    "ps = torch.exp(log_ps)\n",
    "\n",
    "# The test image and the predicted class probabilities\n",
    "import helper\n",
    "helper.view_classify(img.view(28,28),ps,version='Fashion')\n",
    "print(labels[0])\n",
    "\n",
    "# Set the model back to training mode\n",
    "model.train()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "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": "3.6.9"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}