PyTorch Tutorial-DataLoaders.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "cab4753a",
   "metadata": {},
   "source": [
    "# PyTorch Tutorial, DataLoaders:\n",
    "#### DataLoaders get Datasets and convert them into iterables\n",
    "- Download the **iris** dataset, then apply **TensorDataset**\n",
    "    - See previous post for more details on this step\n",
    "- Apply **DataLoader** over the tensor Dataset, to make dataset iterable.\n",
    "    - Using DataLoader with iter() and next() methods\n",
    "    - Using DataLoader in a for-loop\n",
    "- Download the **MNIST hand-written digit** dataset and apply DataLoader to it.\n",
    "    - Then, get and display a minibatch of the MNIST dataset using iter() and next() methods.\n",
    "\n",
    "https://github.com/ostad-ai/PyTorch-Tutorial"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "75354f9d",
   "metadata": {},
   "outputs": [],
   "source": [
    "# importing the necessary modules\n",
    "import torch\n",
    "from torchvision import datasets,transforms\n",
    "from torch.utils.data import TensorDataset,DataLoader\n",
    "from matplotlib import pyplot as plt\n",
    "from urllib import request"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1c5e5ce0",
   "metadata": {},
   "source": [
    "Read the following Jupyter Notebook for more info on the iris dataset: <br>\n",
    "https://raw.githubusercontent.com/ostad-ai/Machine-Learning/main/data%20visualization-Iris%20dataset%20with%20Blender%20API.ipynb\n",
    "<br> Loading the **iris** dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "1624f1be",
   "metadata": {},
   "outputs": [],
   "source": [
    "file_url='https://raw.githubusercontent.com/ostad-ai/Machine-Learning/main/iris.csv'\n",
    "with request.urlopen(file_url) as file:\n",
    "    iris=file.read().decode('utf-8').splitlines()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "71f0a8d0",
   "metadata": {},
   "source": [
    "Separating **features** and **labels** from the iris dataset in the form data pair (features,labels).\n",
    "<br> Then, applying **TensorDataset** to get in the form of **Dataset** in PyTorch (see previous post in the repository for more details).\n",
    "<br> The **iris** dataset has **four** features and **three** classes. The dataset has **150** samples.\n",
    "<br>Reminder: The class **Dataset** in **PyTorch** holds the features and labels of samples of the given dataset in form of (x,y) pairs."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "5b52b033",
   "metadata": {},
   "outputs": [],
   "source": [
    "def iris_features_labels(dataset):\n",
    "    header=[]; rows=[]\n",
    "    for line in dataset:\n",
    "        if line.strip():\n",
    "            if not header: header=line.split(',')\n",
    "            else: rows.append(line.split(','))   \n",
    "    features=[]; labels=[]\n",
    "    classes=list(set([row[4] for row in rows]))\n",
    "    for row in rows:\n",
    "        features.append([float(item) for item in row[:4]])\n",
    "        labels.append(classes.index(row[4]))\n",
    "    return features,labels,classes,header\n",
    "features,labels,classes,header=iris_features_labels(iris)\n",
    "# converting to tensors\n",
    "features_tn,labels_tn=torch.tensor(features),torch.tensor(labels)\n",
    "#converting to tensors\n",
    "xy=TensorDataset(features_tn,labels_tn)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "80ff81e9",
   "metadata": {},
   "source": [
    "A complete pass over the entire dataset is called an **epoch**.\n",
    "In each epoch, we feed our neural network (deep learning) with small groups of the dataset. We call such small groups of dataset **minibatches**. \n",
    "<br>After each minibatch, weights of our neural network are updated.\n",
    "<br>To reduce **overfitting**, at the end of each epoch ,we should *reshuffle* the dataset.\n",
    "<br> **DataLoader** gets the dataset and converts the dataset into an iterable. DataLoader is an iterator class which supports iter() and next() methods. By these methods, we can iterate over the dataset."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "fa6323f5",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Features in a minibatch of size 7:\n",
      "tensor([[6.1000, 2.8000, 4.7000, 1.2000],\n",
      "        [5.6000, 3.0000, 4.5000, 1.5000],\n",
      "        [4.9000, 3.1000, 1.5000, 0.1000],\n",
      "        [5.7000, 3.0000, 4.2000, 1.2000],\n",
      "        [5.5000, 4.2000, 1.4000, 0.2000],\n",
      "        [5.8000, 2.7000, 3.9000, 1.2000],\n",
      "        [5.6000, 2.9000, 3.6000, 1.3000]])\n",
      "Labels of the minibatch of size 7:\n",
      "tensor([1, 1, 2, 1, 2, 1, 1])\n"
     ]
    }
   ],
   "source": [
    "minibatch_size=7\n",
    "xy_dataloader=DataLoader(xy,batch_size=minibatch_size,shuffle=True)\n",
    "xy_iter=iter(xy_dataloader)\n",
    "xs0,ys0=next(xy_iter)\n",
    "print(f'Features in a minibatch of size {minibatch_size}:\\n{xs0}')\n",
    "print(f'Labels of the minibatch of size {minibatch_size}:\\n{ys0}')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3edc2b84",
   "metadata": {},
   "source": [
    "We can also iterate over DataLoader in a **for-loop**, as shown below. \n",
    "<br>In the example below, we just show three minibatches.\n",
    "<br> Hint: The parameter batch_size of DataLoader is the size of a minibatch. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "105f6729",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Each minibatch is composed of 4 pairs (feature,label)\n",
      "Each feature has four components, and each label has one component\n",
      "minibatch: 0\n",
      "tensor([[4.9000, 2.5000, 4.5000, 1.7000],\n",
      "        [6.2000, 2.9000, 4.3000, 1.3000],\n",
      "        [4.7000, 3.2000, 1.6000, 0.2000],\n",
      "        [6.0000, 2.2000, 4.0000, 1.0000]]) tensor([0, 1, 2, 1])\n",
      "minibatch: 1\n",
      "tensor([[7.0000, 3.2000, 4.7000, 1.4000],\n",
      "        [5.0000, 3.6000, 1.4000, 0.2000],\n",
      "        [6.7000, 3.1000, 4.4000, 1.4000],\n",
      "        [7.7000, 3.8000, 6.7000, 2.2000]]) tensor([1, 2, 1, 0])\n",
      "minibatch: 2\n",
      "tensor([[5.2000, 2.7000, 3.9000, 1.4000],\n",
      "        [4.3000, 3.0000, 1.1000, 0.1000],\n",
      "        [5.1000, 3.3000, 1.7000, 0.5000],\n",
      "        [5.7000, 4.4000, 1.5000, 0.4000]]) tensor([1, 2, 2, 2])\n"
     ]
    }
   ],
   "source": [
    "minibatch_size=4\n",
    "xy_dataloader2=DataLoader(xy,batch_size=minibatch_size,shuffle=True)\n",
    "N2show=3; n2show=0\n",
    "print(f'Each minibatch is composed of {minibatch_size} pairs (feature,label)')\n",
    "print('Each feature has four components, and each label has one component')\n",
    "#using Dataloader in a for-loop\n",
    "for x,y in xy_dataloader2:\n",
    "    print(f'minibatch: {n2show}')\n",
    "    print(x,y)\n",
    "    n2show+=1\n",
    "    if n2show>=N2show:\n",
    "        break"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "64e403f9",
   "metadata": {},
   "source": [
    "Let's download the **MNIST** dataset which contains handwritten digits in gray-level images, each of size 28*28"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "555766cf",
   "metadata": {},
   "outputs": [],
   "source": [
    "# here, we download the training set of MNIST\n",
    "mnist=datasets.MNIST(root='./MNIST',train=True, download=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "50aeb465",
   "metadata": {},
   "source": [
    "Display a sample of MNIST data pair: (image,label)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "6d63d840",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "type: <class 'PIL.Image.Image'>, width:28 height:28\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# the original images in MNIST are in the format PIL\n",
    "mnist.transform=None\n",
    "image,label=mnist[10]\n",
    "print(f'type: {type(image)}, width:{image.width} height:{image.height}')\n",
    "plt.imshow(image,cmap='gray')\n",
    "plt.title(f'label: {label}')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6935d7ae",
   "metadata": {},
   "source": [
    "Get a minibatch of MNIST dataset, after using ToTensor() to it, employing DataLoader, and showing the images with pyplot"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "352b9078",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1000x1000 with 20 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# images in mnist are converted to tensor as we iterate\n",
    "mnist.transform=transforms.ToTensor()\n",
    "mnist_bs=20 # size of each minibatch\n",
    "mnist_dataloader=DataLoader(mnist,batch_size=mnist_bs,shuffle=True)\n",
    "minibatch_0=next(iter(mnist_dataloader))\n",
    "images,labels=minibatch_0\n",
    "plt.figure(figsize=(10,10))\n",
    "for i in range(mnist_bs):\n",
    "    plt.subplot(4,5,i+1)\n",
    "    plt.imshow(images[i].permute(1,2,0),cmap='gray')\n",
    "    plt.title(f'label:{labels[i]}')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "f8ec45d3",
   "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": "3.11.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}