MNIST Feed Forward.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "f151f3e9",
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy\n",
    "import torch\n",
    "import torch.nn as nn\n",
    "import torchvision.datasets as datasets\n",
    "import torchvision.transforms as transforms"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "12414417",
   "metadata": {},
   "outputs": [],
   "source": [
    "# importing the datasets\n",
    "training_dataset = datasets.MNIST(root = './data',\n",
    "                                  train = True,\n",
    "                                  transform = transforms.ToTensor(),\n",
    "                                  download = True)\n",
    "\n",
    "test_dataset = datasets.MNIST(root = './data',\n",
    "                              train = False,\n",
    "                              transform = transforms.ToTensor(),\n",
    "                              download = True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "fc2dd756",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "torch.Size([60000, 28, 28])"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "training_dataset.data.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "5510715f",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "torch.Size([28, 28])"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "image = training_dataset.data[0]\n",
    "image.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "8ed8b607",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "\n",
    "# Visualizing training images\n",
    "index = 7\n",
    "image = training_dataset.data[index]\n",
    "label = training_dataset.targets[index]\n",
    "plt.imshow(image, cmap='gray')\n",
    "plt.title(label.item())\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "17ce5c1b",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Spliting the training dataset into batches\n",
    "training_dataloader = torch.utils.data.DataLoader(dataset = training_dataset,\n",
    "                                                  batch_size = 100,\n",
    "                                                  shuffle = True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "e0c16343",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of batches: 600\n",
      "Each batch has: 100 images and 100 labels.\n",
      "Batch input shape: torch.Size([100, 1, 28, 28])\n",
      "Batch output shape: torch.Size([100])\n",
      "Image shape: torch.Size([1, 28, 28])\n",
      "Flatten shape: torch.Size([100, 784])\n"
     ]
    }
   ],
   "source": [
    "for images, labels in training_dataloader:\n",
    "    print(\"Number of batches:\", len(training_dataset)//images.shape[0])\n",
    "    print(f\"Each batch has: {images.shape[0]} images and {labels.shape[0]} labels.\")\n",
    "    print(\"Batch input shape:\", images.shape)\n",
    "    print(\"Batch output shape:\", labels.shape)\n",
    "    print(\"Image shape:\", images[0].shape)\n",
    "    print(\"Flatten shape:\", images.view(-1, 784).shape)\n",
    "    break"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "7869c727",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Defining the parameters\n",
    "input_dim = 784      # 28 * 28\n",
    "hidden_1_dim = 400   # approx. = (input_dim+output_dim)/2\n",
    "hidden_2_dim = 400   # approx. = (input_dim+output_dim)/2\n",
    "output_dim = 10      # 10 digits therfore 10 classes\n",
    "lr = 0.001\n",
    "epochs = 10"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "4964b15a",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Defining the network\n",
    "class Model(nn.Module):\n",
    "    def __init__(self, input_dim, hidden_1_dim, hidden_2_dim, output_dim):\n",
    "        super().__init__()\n",
    "        self.fc1 = nn.Linear(input_dim, hidden_1_dim)\n",
    "        self.fc2 = nn.Linear(hidden_1_dim, hidden_2_dim)\n",
    "        self.fc3 = nn.Linear(hidden_2_dim, output_dim)\n",
    "        self.relu = nn.ReLU()\n",
    "    \n",
    "    def forward(self, X):\n",
    "        out = self.fc1(X)\n",
    "        out = self.relu(out)\n",
    "        out = self.fc2(out)\n",
    "        out = self.relu(out)\n",
    "        out = self.fc3(out)\n",
    "        return out"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "c973c408",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Building the network\n",
    "model = Model(input_dim, hidden_1_dim, hidden_2_dim, output_dim)\n",
    "\n",
    "# loss function\n",
    "loss_fn = nn.CrossEntropyLoss()\n",
    "\n",
    "# optimizer function\n",
    "optimizer = torch.optim.Adam(model.parameters(), lr=lr)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "7f108214",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<bound method Module.parameters of Model(\n",
       "  (fc1): Linear(in_features=784, out_features=400, bias=True)\n",
       "  (fc2): Linear(in_features=400, out_features=400, bias=True)\n",
       "  (fc3): Linear(in_features=400, out_features=10, bias=True)\n",
       "  (relu): ReLU()\n",
       ")>"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.parameters"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "2fa2b495",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch: 1/10, iteration: 100/600, loss: 0.259,                   accuracy: 81.21%\n",
      "Epoch: 1/10, iteration: 200/600, loss: 0.197,                   accuracy: 86.80%\n",
      "Epoch: 1/10, iteration: 300/600, loss: 0.101,                   accuracy: 89.00%\n",
      "Epoch: 1/10, iteration: 400/600, loss: 0.118,                   accuracy: 90.46%\n",
      "Epoch: 1/10, iteration: 500/600, loss: 0.097,                   accuracy: 91.50%\n",
      "Epoch: 1/10, iteration: 600/600, loss: 0.187,                   accuracy: 92.18%\n",
      "Epoch: 2/10, iteration: 100/600, loss: 0.093,                   accuracy: 96.85%\n",
      "Epoch: 2/10, iteration: 200/600, loss: 0.074,                   accuracy: 96.93%\n",
      "Epoch: 2/10, iteration: 300/600, loss: 0.091,                   accuracy: 96.71%\n",
      "Epoch: 2/10, iteration: 400/600, loss: 0.136,                   accuracy: 96.86%\n",
      "Epoch: 2/10, iteration: 500/600, loss: 0.108,                   accuracy: 96.98%\n",
      "Epoch: 2/10, iteration: 600/600, loss: 0.110,                   accuracy: 97.06%\n",
      "Epoch: 3/10, iteration: 100/600, loss: 0.132,                   accuracy: 98.30%\n",
      "Epoch: 3/10, iteration: 200/600, loss: 0.089,                   accuracy: 98.09%\n",
      "Epoch: 3/10, iteration: 300/600, loss: 0.034,                   accuracy: 97.95%\n",
      "Epoch: 3/10, iteration: 400/600, loss: 0.059,                   accuracy: 97.97%\n",
      "Epoch: 3/10, iteration: 500/600, loss: 0.021,                   accuracy: 98.02%\n",
      "Epoch: 3/10, iteration: 600/600, loss: 0.057,                   accuracy: 97.99%\n",
      "Epoch: 4/10, iteration: 100/600, loss: 0.030,                   accuracy: 98.68%\n",
      "Epoch: 4/10, iteration: 200/600, loss: 0.006,                   accuracy: 98.64%\n",
      "Epoch: 4/10, iteration: 300/600, loss: 0.051,                   accuracy: 98.57%\n",
      "Epoch: 4/10, iteration: 400/600, loss: 0.058,                   accuracy: 98.57%\n",
      "Epoch: 4/10, iteration: 500/600, loss: 0.059,                   accuracy: 98.54%\n",
      "Epoch: 4/10, iteration: 600/600, loss: 0.040,                   accuracy: 98.56%\n",
      "Epoch: 5/10, iteration: 100/600, loss: 0.012,                   accuracy: 99.31%\n",
      "Epoch: 5/10, iteration: 200/600, loss: 0.025,                   accuracy: 99.15%\n",
      "Epoch: 5/10, iteration: 300/600, loss: 0.035,                   accuracy: 99.02%\n",
      "Epoch: 5/10, iteration: 400/600, loss: 0.030,                   accuracy: 98.94%\n",
      "Epoch: 5/10, iteration: 500/600, loss: 0.015,                   accuracy: 98.89%\n",
      "Epoch: 5/10, iteration: 600/600, loss: 0.011,                   accuracy: 98.90%\n",
      "Epoch: 6/10, iteration: 100/600, loss: 0.022,                   accuracy: 99.36%\n",
      "Epoch: 6/10, iteration: 200/600, loss: 0.007,                   accuracy: 99.29%\n",
      "Epoch: 6/10, iteration: 300/600, loss: 0.047,                   accuracy: 99.25%\n",
      "Epoch: 6/10, iteration: 400/600, loss: 0.054,                   accuracy: 99.19%\n",
      "Epoch: 6/10, iteration: 500/600, loss: 0.093,                   accuracy: 99.17%\n",
      "Epoch: 6/10, iteration: 600/600, loss: 0.010,                   accuracy: 99.12%\n",
      "Epoch: 7/10, iteration: 100/600, loss: 0.015,                   accuracy: 99.40%\n",
      "Epoch: 7/10, iteration: 200/600, loss: 0.042,                   accuracy: 99.33%\n",
      "Epoch: 7/10, iteration: 300/600, loss: 0.005,                   accuracy: 99.32%\n",
      "Epoch: 7/10, iteration: 400/600, loss: 0.040,                   accuracy: 99.29%\n",
      "Epoch: 7/10, iteration: 500/600, loss: 0.043,                   accuracy: 99.26%\n",
      "Epoch: 7/10, iteration: 600/600, loss: 0.005,                   accuracy: 99.26%\n",
      "Epoch: 8/10, iteration: 100/600, loss: 0.001,                   accuracy: 99.63%\n",
      "Epoch: 8/10, iteration: 200/600, loss: 0.017,                   accuracy: 99.47%\n",
      "Epoch: 8/10, iteration: 300/600, loss: 0.032,                   accuracy: 99.43%\n",
      "Epoch: 8/10, iteration: 400/600, loss: 0.018,                   accuracy: 99.44%\n",
      "Epoch: 8/10, iteration: 500/600, loss: 0.002,                   accuracy: 99.41%\n",
      "Epoch: 8/10, iteration: 600/600, loss: 0.035,                   accuracy: 99.40%\n",
      "Epoch: 9/10, iteration: 100/600, loss: 0.002,                   accuracy: 99.64%\n",
      "Epoch: 9/10, iteration: 200/600, loss: 0.035,                   accuracy: 99.49%\n",
      "Epoch: 9/10, iteration: 300/600, loss: 0.004,                   accuracy: 99.48%\n",
      "Epoch: 9/10, iteration: 400/600, loss: 0.005,                   accuracy: 99.49%\n",
      "Epoch: 9/10, iteration: 500/600, loss: 0.006,                   accuracy: 99.47%\n",
      "Epoch: 9/10, iteration: 600/600, loss: 0.001,                   accuracy: 99.47%\n",
      "Epoch: 10/10, iteration: 100/600, loss: 0.001,                   accuracy: 99.50%\n",
      "Epoch: 10/10, iteration: 200/600, loss: 0.000,                   accuracy: 99.65%\n",
      "Epoch: 10/10, iteration: 300/600, loss: 0.005,                   accuracy: 99.63%\n",
      "Epoch: 10/10, iteration: 400/600, loss: 0.000,                   accuracy: 99.62%\n",
      "Epoch: 10/10, iteration: 500/600, loss: 0.001,                   accuracy: 99.56%\n",
      "Epoch: 10/10, iteration: 600/600, loss: 0.005,                   accuracy: 99.52%\n"
     ]
    }
   ],
   "source": [
    "# Training the network\n",
    "for epoch in range(epochs):\n",
    "    total_train = 0\n",
    "    correct_train = 0\n",
    "    for i, (images, labels) in enumerate(training_dataloader):\n",
    "        images = images.view(-1, 784)\n",
    "        \n",
    "        outputs = model(images)\n",
    "        _, predictions = torch.max(outputs, 1)\n",
    "        total_train += len(images)\n",
    "        correct_train += (predictions == labels).sum()\n",
    "        \n",
    "        loss = loss_fn(outputs, labels)\n",
    "        optimizer.zero_grad()\n",
    "        loss.backward()\n",
    "        optimizer.step()\n",
    "        \n",
    "        if (i+1)%100 == 0:\n",
    "            print(f\"Epoch: {epoch+1}/{epochs}, iteration: {i+1}/{len(training_dataset)//images.shape[0]}, loss: {loss:.3f}, \\\n",
    "                  accuracy: {(correct_train/total_train)*100:.2f}%\")\n",
    "        "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "54e25a4e",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 97.26%\n"
     ]
    }
   ],
   "source": [
    "# Calculating testing accuracy\n",
    "test_loader = torch.utils.data.DataLoader(dataset=test_dataset, batch_size=100)\n",
    "\n",
    "correct_preds = 0\n",
    "for images, labels in test_loader:\n",
    "    images = images.view(-1, 784)\n",
    "    _, predictions = torch.max(model(images), 1)\n",
    "    correct_preds += (predictions == labels).sum()\n",
    "\n",
    "print(\"Accuracy: {:.2f}%\".format(correct_preds/len(test_dataset)*100.0))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "404a4a70",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Visualzing the prediction\n",
    "index = 100\n",
    "_, prediction = torch.max(model(test_dataset.data[index].float().view(-1, 784)), 1)\n",
    "plt.imshow(test_dataset.data[index], cmap='gray')\n",
    "plt.title(f'Label: {test_dataset.targets[index]} | Prediction: {test_dataset.classes[prediction]}')\n",
    "plt.show()"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.9.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}