Generalization-checkpoint.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "05f3a1b8",
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch\n",
    "import torch.nn as nn\n",
    "import torch.nn.functional as F\n",
    "from torch.utils.data import DataLoader\n",
    "from torchvision import datasets, transforms\n",
    "\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "f63983f6",
   "metadata": {},
   "outputs": [],
   "source": [
    "transform = transforms.ToTensor()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "8964a97f",
   "metadata": {},
   "outputs": [],
   "source": [
    "train_data_set = datasets.MNIST(root='C:\\\\Users\\\\varsh\\\\dataset\\\\mnist_train', train=True, download=True, transform=transform)\n",
    "test_data_set = datasets.MNIST(root='C:\\\\Users\\\\varsh\\\\dataset\\\\mnist_test', train=False, download=True, transform=transform)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "a49f41a9",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "60000"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(train_data_set)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "3edb4486",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Dataset MNIST\n",
       "    Number of datapoints: 10000\n",
       "    Root location: C:\\Users\\varsh\\dataset\\mnist_test\n",
       "    Split: Test\n",
       "    StandardTransform\n",
       "Transform: ToTensor()"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_data_set"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "92fa3e38",
   "metadata": {},
   "outputs": [],
   "source": [
    "rand_labels_train = torch.tensor(np.random.randint(0,10, (len(train_data_set)),))\n",
    "train_data_set.targets = rand_labels_train\n",
    "rand_labels_test = torch.tensor(np.random.randint(0,10, (len(test_data_set)),))\n",
    "test_data_set.targets = rand_labels_test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "55477839",
   "metadata": {},
   "outputs": [],
   "source": [
    "train_loader = DataLoader(train_data_set, batch_size=50, shuffle=True)\n",
    "test_loader = DataLoader(test_data_set, batch_size=50, shuffle=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "42b1c1ab",
   "metadata": {},
   "outputs": [],
   "source": [
    "class DNModel(nn.Module):\n",
    "    def __init__(self, i=784, z1=120, z2=120,z3=16, o=10):\n",
    "        super().__init__()\n",
    "        self.fc1 = nn.Linear(i,z1)    \n",
    "        self.fc2 = nn.Linear(z1, z2)\n",
    "        self.fc3 = nn.Linear(z2, z3)\n",
    "        self.out = nn.Linear(z3, o)\n",
    "        \n",
    "        \n",
    "    def forward(self, k):\n",
    "        k = F.relu(self.fc1(k))\n",
    "        k = F.relu(self.fc2(k))\n",
    "        k = F.relu(self.fc3(k))\n",
    "        k = self.out(k)\n",
    "        return k"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "cec446f9",
   "metadata": {},
   "outputs": [],
   "source": [
    "dn_model=DNModel()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "9ca38b6a",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "DNModel(\n",
       "  (fc1): Linear(in_features=784, out_features=120, bias=True)\n",
       "  (fc2): Linear(in_features=120, out_features=120, bias=True)\n",
       "  (fc3): Linear(in_features=120, out_features=16, bias=True)\n",
       "  (out): Linear(in_features=16, out_features=10, bias=True)\n",
       ")"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dn_model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "380efac9",
   "metadata": {},
   "outputs": [],
   "source": [
    "CriterionLoss = nn.CrossEntropyLoss()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "6f98564a",
   "metadata": {},
   "outputs": [],
   "source": [
    "optim = torch.optim.Adam(dn_model.parameters(), lr=0.001)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "85b0ea14",
   "metadata": {},
   "outputs": [],
   "source": [
    "epochs = 10\n",
    "train_losses = []\n",
    "test_losses = []\n",
    "\n",
    "for i in range(epochs):\n",
    "    \n",
    "    #training the batches\n",
    "    for b, (Xtrain, ytrain) in enumerate(train_loader):\n",
    "        \n",
    "        # Applying the model\n",
    "        ypred = dn_model(Xtrain.view(50, -1))  # Here we flatten X_train\n",
    "        loss = CriterionLoss(ypred, ytrain)\n",
    " \n",
    "        # Updating the parameters\n",
    "        optim.zero_grad()\n",
    "        loss.backward()\n",
    "        optim.step()\n",
    "        \n",
    "     \n",
    "    # Appending train loss  for the epoch\n",
    "    train_losses.append(loss)\n",
    "        \n",
    "    # testing the batches\n",
    "    with torch.no_grad():\n",
    "        for b, (Xtest, ytest) in enumerate(test_loader):\n",
    "\n",
    "            # Applying  the model\n",
    "            yval = dn_model(Xtest.view(50, -1)) \n",
    "    \n",
    "    # Appending test loss for the epoch\n",
    "    loss = CriterionLoss(yval, ytest)\n",
    "    test_losses.append(loss)\n",
    "    \n",
    "        "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "0e99200d",
   "metadata": {},
   "outputs": [],
   "source": [
    "train_losses=torch.tensor(train_losses)\n",
    "test_losses=torch.tensor(test_losses)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "b167e8e6",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Visulaize losses for for MNIST dataset\n",
    "plt.plot(range(epochs), train_losses, 'r', label='Train')\n",
    "plt.plot(range(epochs), test_losses, 'b', label='Test')\n",
    "plt.title(\"Loss\")\n",
    "plt.xlabel(\"Epochs\")\n",
    "plt.ylabel(\"Loss\")\n",
    "plt.legend(loc=\"upper right\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "01e2ff30",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "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": "3.9.13"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}