PyTorch Concepts(Validation set Concept).ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "#we have worked with the fashion mnist data now we can easily overfit the data \n",
    "#to stop that we need to make a validation set and stop training where the validation error start incerasing \n",
    "\n",
    "#previous model\n",
    "\n",
    "import torch\n",
    "from torchvision import datasets, transforms\n",
    "import helper\n",
    "\n",
    "# Define a transform to normalize the data\n",
    "transform = transforms.Compose([transforms.ToTensor(),\n",
    "                                transforms.Normalize((0.5, 0.5, 0.5), (0.5, 0.5, 0.5))])\n",
    "# Download and load the training data\n",
    "trainset = datasets.FashionMNIST('~/.pytorch/F_MNIST_data/', download=False, 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=False, train=False, transform=transform)\n",
    "testloader = torch.utils.data.DataLoader(testset, batch_size=64, shuffle=True)\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "from torch import nn,optim\n",
    "import torch.nn.functional as F"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = nn.Sequential(\n",
    "                    nn.Linear(784,128),\n",
    "                    nn.ReLU(),\n",
    "                    nn.Linear(128,64),\n",
    "                    nn.ReLU(),\n",
    "                    nn.Linear(64,10),\n",
    "                    nn.LogSoftmax(dim=1)\n",
    ")\n",
    "#defining the loss\n",
    "criterion = nn.NLLLoss()\n",
    "optimizer = optim.Adam(model.parameters(),lr=0.003)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "%config IPCompleter.greedy=True"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "torch.Size([64, 10])\n"
     ]
    }
   ],
   "source": [
    "images,labels = next(iter(testloader))\n",
    "images = images.view(64,784)\n",
    "ps = torch.exp(model(images))\n",
    "print(ps.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "torch.Size([64, 5])\n",
      "tensor([[0.1147, 0.1103, 0.1055, 0.1033, 0.1007],\n",
      "        [0.1153, 0.1110, 0.1053, 0.1016, 0.0996],\n",
      "        [0.1124, 0.1084, 0.1072, 0.1025, 0.1016],\n",
      "        [0.1140, 0.1111, 0.1020, 0.1019, 0.1009],\n",
      "        [0.1144, 0.1091, 0.1059, 0.0996, 0.0992],\n",
      "        [0.1138, 0.1091, 0.1045, 0.1027, 0.1026],\n",
      "        [0.1105, 0.1100, 0.1074, 0.1030, 0.1029],\n",
      "        [0.1139, 0.1072, 0.1064, 0.1030, 0.1013],\n",
      "        [0.1147, 0.1124, 0.1058, 0.1050, 0.1018],\n",
      "        [0.1100, 0.1093, 0.1045, 0.1030, 0.1019]], grad_fn=<SliceBackward>)\n",
      "torch.Size([64, 5])\n",
      "tensor([[4, 1, 8, 3, 6],\n",
      "        [1, 4, 0, 3, 8],\n",
      "        [3, 1, 4, 0, 6],\n",
      "        [4, 1, 7, 8, 3],\n",
      "        [1, 4, 3, 7, 0],\n",
      "        [1, 4, 3, 8, 7],\n",
      "        [4, 1, 8, 6, 3],\n",
      "        [3, 1, 4, 6, 7],\n",
      "        [1, 4, 8, 6, 3],\n",
      "        [1, 3, 6, 4, 0]])\n"
     ]
    }
   ],
   "source": [
    "#we will use the method ps.topk which will return the k highest value of the classes we need\n",
    "top_p,top_class = ps.topk(5,dim=1)\n",
    "print(top_p.shape)\n",
    "print(top_p[:10,:])\n",
    "print(top_class.shape)\n",
    "print(top_class[:10,:])\n",
    "#we see that all got nearly equal probability as our model is not trained yet"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "top_p,top_class = ps.topk(1,dim=1)\n",
    "equals = top_class == labels.view(*top_class.shape)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 15.625%\n"
     ]
    }
   ],
   "source": [
    "#we have to convert the data type because equal has a datatype of byte tensor but we need float tensor to find the mean\n",
    "accuracy = torch.mean(equals.type(torch.FloatTensor))\n",
    "print(f'Accuracy: {accuracy.item()*100}%')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch: 1/30..  Training Loss: 0.267..  Test Loss: 0.368..  Test Accuracy: 0.869\n",
      "Epoch: 2/30..  Training Loss: 0.259..  Test Loss: 0.366..  Test Accuracy: 0.877\n",
      "Epoch: 3/30..  Training Loss: 0.253..  Test Loss: 0.362..  Test Accuracy: 0.874\n",
      "Epoch: 4/30..  Training Loss: 0.245..  Test Loss: 0.416..  Test Accuracy: 0.866\n",
      "Epoch: 5/30..  Training Loss: 0.239..  Test Loss: 0.388..  Test Accuracy: 0.871\n",
      "Epoch: 6/30..  Training Loss: 0.240..  Test Loss: 0.373..  Test Accuracy: 0.877\n",
      "Epoch: 7/30..  Training Loss: 0.231..  Test Loss: 0.395..  Test Accuracy: 0.874\n",
      "Epoch: 8/30..  Training Loss: 0.228..  Test Loss: 0.375..  Test Accuracy: 0.882\n",
      "Epoch: 9/30..  Training Loss: 0.221..  Test Loss: 0.404..  Test Accuracy: 0.875\n",
      "Epoch: 10/30..  Training Loss: 0.220..  Test Loss: 0.383..  Test Accuracy: 0.880\n",
      "Epoch: 11/30..  Training Loss: 0.213..  Test Loss: 0.376..  Test Accuracy: 0.877\n",
      "Epoch: 12/30..  Training Loss: 0.214..  Test Loss: 0.402..  Test Accuracy: 0.876\n",
      "Epoch: 13/30..  Training Loss: 0.206..  Test Loss: 0.424..  Test Accuracy: 0.878\n",
      "Epoch: 14/30..  Training Loss: 0.209..  Test Loss: 0.420..  Test Accuracy: 0.874\n",
      "Epoch: 15/30..  Training Loss: 0.201..  Test Loss: 0.400..  Test Accuracy: 0.876\n",
      "Epoch: 16/30..  Training Loss: 0.197..  Test Loss: 0.420..  Test Accuracy: 0.875\n",
      "Epoch: 17/30..  Training Loss: 0.200..  Test Loss: 0.397..  Test Accuracy: 0.879\n",
      "Epoch: 18/30..  Training Loss: 0.192..  Test Loss: 0.460..  Test Accuracy: 0.876\n",
      "Epoch: 19/30..  Training Loss: 0.195..  Test Loss: 0.444..  Test Accuracy: 0.881\n",
      "Epoch: 20/30..  Training Loss: 0.187..  Test Loss: 0.459..  Test Accuracy: 0.874\n",
      "Epoch: 21/30..  Training Loss: 0.183..  Test Loss: 0.427..  Test Accuracy: 0.878\n",
      "Epoch: 22/30..  Training Loss: 0.184..  Test Loss: 0.448..  Test Accuracy: 0.875\n",
      "Epoch: 23/30..  Training Loss: 0.181..  Test Loss: 0.448..  Test Accuracy: 0.878\n",
      "Epoch: 24/30..  Training Loss: 0.175..  Test Loss: 0.465..  Test Accuracy: 0.876\n",
      "Epoch: 25/30..  Training Loss: 0.180..  Test Loss: 0.461..  Test Accuracy: 0.874\n",
      "Epoch: 26/30..  Training Loss: 0.171..  Test Loss: 0.467..  Test Accuracy: 0.878\n",
      "Epoch: 27/30..  Training Loss: 0.171..  Test Loss: 0.471..  Test Accuracy: 0.881\n",
      "Epoch: 28/30..  Training Loss: 0.174..  Test Loss: 0.435..  Test Accuracy: 0.883\n",
      "Epoch: 29/30..  Training Loss: 0.169..  Test Loss: 0.461..  Test Accuracy: 0.877\n",
      "Epoch: 30/30..  Training Loss: 0.166..  Test Loss: 0.465..  Test Accuracy: 0.882\n"
     ]
    }
   ],
   "source": [
    "#trainig the model\n",
    "epochs = 30\n",
    "steps = 0\n",
    "train_losses, test_losses = [], []\n",
    "for e in range(epochs):\n",
    "    running_loss = 0\n",
    "    for images,labels in trainloader:\n",
    "        optimizer.zero_grad()\n",
    "        images = images.view(images.shape[0],-1)\n",
    "        log_ps = model(images)\n",
    "        loss=criterion(log_ps,labels)\n",
    "        loss.backward()\n",
    "        optimizer.step()\n",
    "        running_loss+=loss.item()\n",
    "    else:\n",
    "        test_loss = 0\n",
    "        accuracy = 0\n",
    "        #turning off the gradients will save memory and time\n",
    "        with torch.no_grad():\n",
    "            for images,labels in testloader:\n",
    "                images = images.view(images.shape[0],-1)\n",
    "                log_ps = model(images)\n",
    "                test_loss += criterion(log_ps,labels)\n",
    "                \n",
    "                ps = torch.exp(log_ps)\n",
    "                top_p,top_class = ps.topk(1,dim=1)\n",
    "                equals = top_class == labels.view(*top_class.shape)\n",
    "                accuracy += torch.mean(equals.type(torch.FloatTensor))\n",
    "        train_losses.append(running_loss/len(trainloader))\n",
    "        test_losses.append(test_loss/len(testloader))\n",
    "\n",
    "        print(\"Epoch: {}/{}.. \".format(e+1, epochs),\n",
    "              \"Training Loss: {:.3f}.. \".format(running_loss/len(trainloader)),\n",
    "              \"Test Loss: {:.3f}.. \".format(test_loss/len(testloader)),\n",
    "              \"Test Accuracy: {:.3f}\".format(accuracy/len(testloader)))\n",
    "                \n",
    "        "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x7f4b1f824c50>"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f4b5fd72208>"
      ]
     },
     "metadata": {
      "image/png": {
       "height": 250,
       "width": 380
      }
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "%matplotlib inline\n",
    "%config InlineBackend.figure_format = 'retina'\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": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [],
   "source": [
    "#to avoid overfitting we can use the dropout layer\n",
    "#which will drop some gradient in calculation of gradients thus helping us overall\n",
    "model = nn.Sequential(\n",
    "                    nn.Linear(784,128),\n",
    "                    nn.ReLU(),\n",
    "                    nn.Dropout(),\n",
    "                    nn.Linear(128,64),\n",
    "                    nn.ReLU(),\n",
    "                    nn.Dropout(),\n",
    "                    nn.Linear(64,10),\n",
    "                    nn.LogSoftmax(dim=1)\n",
    ")\n",
    "#defining the loss\n",
    "criterion = nn.NLLLoss()\n",
    "optimizer = optim.Adam(model.parameters(),lr=0.003)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch: 1/30..  Training Loss: 0.787..  Test Loss: 0.527..  Test Accuracy: 0.800\n",
      "Epoch: 2/30..  Training Loss: 0.641..  Test Loss: 0.505..  Test Accuracy: 0.807\n",
      "Epoch: 3/30..  Training Loss: 0.610..  Test Loss: 0.468..  Test Accuracy: 0.830\n",
      "Epoch: 4/30..  Training Loss: 0.592..  Test Loss: 0.449..  Test Accuracy: 0.837\n",
      "Epoch: 5/30..  Training Loss: 0.579..  Test Loss: 0.456..  Test Accuracy: 0.833\n",
      "Epoch: 6/30..  Training Loss: 0.573..  Test Loss: 0.461..  Test Accuracy: 0.834\n",
      "Epoch: 7/30..  Training Loss: 0.562..  Test Loss: 0.442..  Test Accuracy: 0.834\n",
      "Epoch: 8/30..  Training Loss: 0.554..  Test Loss: 0.437..  Test Accuracy: 0.839\n",
      "Epoch: 9/30..  Training Loss: 0.552..  Test Loss: 0.445..  Test Accuracy: 0.835\n",
      "Epoch: 10/30..  Training Loss: 0.544..  Test Loss: 0.429..  Test Accuracy: 0.843\n",
      "Epoch: 11/30..  Training Loss: 0.547..  Test Loss: 0.424..  Test Accuracy: 0.844\n",
      "Epoch: 12/30..  Training Loss: 0.534..  Test Loss: 0.429..  Test Accuracy: 0.845\n",
      "Epoch: 13/30..  Training Loss: 0.539..  Test Loss: 0.422..  Test Accuracy: 0.850\n",
      "Epoch: 14/30..  Training Loss: 0.529..  Test Loss: 0.419..  Test Accuracy: 0.848\n",
      "Epoch: 15/30..  Training Loss: 0.526..  Test Loss: 0.414..  Test Accuracy: 0.848\n",
      "Epoch: 16/30..  Training Loss: 0.528..  Test Loss: 0.428..  Test Accuracy: 0.846\n",
      "Epoch: 17/30..  Training Loss: 0.524..  Test Loss: 0.421..  Test Accuracy: 0.846\n",
      "Epoch: 18/30..  Training Loss: 0.524..  Test Loss: 0.419..  Test Accuracy: 0.847\n",
      "Epoch: 19/30..  Training Loss: 0.517..  Test Loss: 0.411..  Test Accuracy: 0.852\n",
      "Epoch: 20/30..  Training Loss: 0.520..  Test Loss: 0.419..  Test Accuracy: 0.850\n",
      "Epoch: 21/30..  Training Loss: 0.517..  Test Loss: 0.435..  Test Accuracy: 0.842\n",
      "Epoch: 22/30..  Training Loss: 0.514..  Test Loss: 0.409..  Test Accuracy: 0.849\n",
      "Epoch: 23/30..  Training Loss: 0.515..  Test Loss: 0.425..  Test Accuracy: 0.850\n",
      "Epoch: 24/30..  Training Loss: 0.508..  Test Loss: 0.425..  Test Accuracy: 0.845\n",
      "Epoch: 25/30..  Training Loss: 0.514..  Test Loss: 0.417..  Test Accuracy: 0.848\n",
      "Epoch: 26/30..  Training Loss: 0.513..  Test Loss: 0.434..  Test Accuracy: 0.835\n",
      "Epoch: 27/30..  Training Loss: 0.504..  Test Loss: 0.426..  Test Accuracy: 0.850\n",
      "Epoch: 28/30..  Training Loss: 0.501..  Test Loss: 0.417..  Test Accuracy: 0.855\n",
      "Epoch: 29/30..  Training Loss: 0.506..  Test Loss: 0.414..  Test Accuracy: 0.849\n",
      "Epoch: 30/30..  Training Loss: 0.507..  Test Loss: 0.410..  Test Accuracy: 0.854\n"
     ]
    }
   ],
   "source": [
    "#we can use model.eval() which turn off the gradients and also set dropout probability to 0\n",
    "train_losses, test_losses = [], []\n",
    "for e in range(epochs):\n",
    "    running_loss = 0\n",
    "    for images, labels in trainloader:\n",
    "        \n",
    "        optimizer.zero_grad()\n",
    "        images = images.view(images.shape[0],-1)\n",
    "        log_ps = model(images)\n",
    "        loss = criterion(log_ps, labels)\n",
    "        loss.backward()\n",
    "        optimizer.step()\n",
    "        \n",
    "        running_loss += loss.item()\n",
    "        \n",
    "    else:\n",
    "        test_loss = 0\n",
    "        accuracy = 0\n",
    "        \n",
    "        # Turn off gradients for validation, saves memory and computations\n",
    "        with torch.no_grad():\n",
    "            model.eval()\n",
    "            for images, labels in testloader:\n",
    "                images = images.view(images.shape[0],-1)\n",
    "                log_ps = model(images)\n",
    "                test_loss += criterion(log_ps, labels)\n",
    "                \n",
    "                ps = torch.exp(log_ps)\n",
    "                top_p, top_class = ps.topk(1, dim=1)\n",
    "                equals = top_class == labels.view(*top_class.shape)\n",
    "                accuracy += torch.mean(equals.type(torch.FloatTensor))\n",
    "        #reseting the effect of eval mode\n",
    "        model.train()\n",
    "        \n",
    "        train_losses.append(running_loss/len(trainloader))\n",
    "        test_losses.append(test_loss/len(testloader))\n",
    "\n",
    "        print(\"Epoch: {}/{}.. \".format(e+1, epochs),\n",
    "              \"Training Loss: {:.3f}.. \".format(train_losses[-1]),\n",
    "              \"Test Loss: {:.3f}.. \".format(test_losses[-1]),\n",
    "              \"Test Accuracy: {:.3f}\".format(accuracy/len(testloader)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x7f4b1f791d30>"
      ]
     },
     "execution_count": 40,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f4b2017c0f0>"
      ]
     },
     "metadata": {
      "image/png": {
       "height": 250,
       "width": 380
      }
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "%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": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f4b1f8bf550>"
      ]
     },
     "metadata": {
      "image/png": {
       "height": 204,
       "width": 423
      }
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#now the model can be used for infrence mode \n",
    "import helper\n",
    "\n",
    "# Test out your network!\n",
    "\n",
    "model.eval()\n",
    "\n",
    "dataiter = iter(testloader)\n",
    "images, labels = dataiter.next()\n",
    "img = images[0]\n",
    "# Convert 2D image to 1D vector\n",
    "img = img.view(1, 784)\n",
    "\n",
    "# Calculate the class probabilities (softmax) for img\n",
    "with torch.no_grad():\n",
    "    output = model.forward(img)\n",
    "\n",
    "ps = torch.exp(output)\n",
    "\n",
    "# Plot the image and probabilities\n",
    "helper.view_classify(img.view(1, 28, 28), ps, version='Fashion')"
   ]
  },
  {
   "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": "3.6.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}