Mastering PyTorch Tensors Fundamentals to Advance Guide.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "d8b3b6d1",
   "metadata": {
    "papermill": {
     "duration": 0.038142,
     "end_time": "2023-10-28T03:01:38.712486",
     "exception": false,
     "start_time": "2023-10-28T03:01:38.674344",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "## **Introduction**\n",
    "\n",
    "**[PyTorch](https://pytorch.org/)** is an open-source machine learning library developed by Facebook's AI Research lab. It’s popular in the deep learning community due to its flexibility and dynamic computation graph, which makes it easier to work with and debug. PyTorch provides a wide range of APIs for both neural networks and tensor computations. It is also known for its pythonic design, making it very straightforward and intuitive for Python developers to learn and implement deep learning algorithms. The extensive community support and plethora of tutorials and documentation available make learning and implementing models using PyTorch an enriching experience.\n",
    "\n",
    "For more detailed information and resources, you can visit the official **[PyTorch website](https://pytorch.org/)**."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "99a81e59",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:38.791307Z",
     "iopub.status.busy": "2023-10-28T03:01:38.790991Z",
     "iopub.status.idle": "2023-10-28T03:01:43.401523Z",
     "shell.execute_reply": "2023-10-28T03:01:43.400537Z"
    },
    "papermill": {
     "duration": 4.652849,
     "end_time": "2023-10-28T03:01:43.403998",
     "exception": false,
     "start_time": "2023-10-28T03:01:38.751149",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2.0.0\n"
     ]
    }
   ],
   "source": [
    "# Importing PyTorch, essential for deep learning tasks.\n",
    "import torch\n",
    "\n",
    "# Importing pandas for data manipulation and analysis.\n",
    "import pandas as pd\n",
    "\n",
    "# Importing numpy for numerical computations.\n",
    "import numpy as np\n",
    "\n",
    "# Importing matplotlib for data visualization.\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "# Printing the version of PyTorch to verify installation.\n",
    "print(torch.__version__)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c63615d8",
   "metadata": {
    "id": "3KdMa8liMKre",
    "papermill": {
     "duration": 0.081999,
     "end_time": "2023-10-28T03:01:43.524720",
     "exception": false,
     "start_time": "2023-10-28T03:01:43.442721",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "## **What is Tensor?**\n",
    "**Tensor:** A tensor is a multi-dimensional array that is the fundamental building block of PyTorch and many other machine learning and deep learning libraries. Tensors are similar to NumPy’s ndarrays, but they also have the additional capability to be used on GPUs for accelerated computing. Tensors are optimized for automatic differentiation (autograd) and are used as the primary data structure to build neural networks in PyTorch.\n",
    "\n",
    "Here are a few key points about tensors:\n",
    "\n",
    "#### **Dimentions**\n",
    "- **0D Tensor:** A 0-dimensional tensor is a single number (scalar).\n",
    "- **1D Tensor:** A 1-dimensional tensor is similar to a vector.\n",
    "- **2D Tensor:** A 2-dimensional tensor is similar to a matrix.\n",
    "- **ND Tensor:** Tensors can have three or more dimensions, allowing them to represent more complex data structures.\n",
    "\n",
    "#### **Data Types**\n",
    "- Tensors can hold data of different data types such as float, int, and boolean.\n",
    "\n",
    "#### **Device Compatibility**\n",
    "- Tensors can be moved onto any device, such as a GPU, to accelerate computing.\n",
    "\n",
    "#### **Autograd**\n",
    "- PyTorch tensors support autograd, meaning that they can automatically compute gradients or derivatives, essential for neural network training."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d8e5d121",
   "metadata": {
    "papermill": {
     "duration": 0.039564,
     "end_time": "2023-10-28T03:01:43.603101",
     "exception": false,
     "start_time": "2023-10-28T03:01:43.563537",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "## **Creating Tensors**"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b8c5e262",
   "metadata": {
    "papermill": {
     "duration": 0.038331,
     "end_time": "2023-10-28T03:01:43.680704",
     "exception": false,
     "start_time": "2023-10-28T03:01:43.642373",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "#### **0D Tensor**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "582e98b5",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:43.760928Z",
     "iopub.status.busy": "2023-10-28T03:01:43.760043Z",
     "iopub.status.idle": "2023-10-28T03:01:43.802134Z",
     "shell.execute_reply": "2023-10-28T03:01:43.801206Z"
    },
    "id": "I45Vpb2XLs4s",
    "outputId": "a63df0b4-5505-4ae7-c407-a05e5c34da69",
    "papermill": {
     "duration": 0.083776,
     "end_time": "2023-10-28T03:01:43.804136",
     "exception": false,
     "start_time": "2023-10-28T03:01:43.720360",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tensor(7)"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Creating a 0-dimensional tensor (scalar) with the value 7 using PyTorch.\n",
    "scalar = torch.tensor(7)\n",
    "scalar"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "16d94fee",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:43.883650Z",
     "iopub.status.busy": "2023-10-28T03:01:43.883347Z",
     "iopub.status.idle": "2023-10-28T03:01:43.887735Z",
     "shell.execute_reply": "2023-10-28T03:01:43.886842Z"
    },
    "id": "_rnwMsB2MkqU",
    "outputId": "aa93db92-8cdb-4268-9d60-015069d34a3e",
    "papermill": {
     "duration": 0.046631,
     "end_time": "2023-10-28T03:01:43.889983",
     "exception": false,
     "start_time": "2023-10-28T03:01:43.843352",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Domention of Scalar is: 0\n"
     ]
    }
   ],
   "source": [
    "# Printing the number of dimensions of the scalar tensor. Expecting 0 because it's a scalar.\n",
    "print(f\"Domention of Scalar is: {scalar.ndim}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "4083aa8c",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:43.966323Z",
     "iopub.status.busy": "2023-10-28T03:01:43.966008Z",
     "iopub.status.idle": "2023-10-28T03:01:43.970487Z",
     "shell.execute_reply": "2023-10-28T03:01:43.969627Z"
    },
    "id": "LLCnynZtM0TY",
    "outputId": "f3ba28d9-8836-4336-92b3-53ab98649c3c",
    "papermill": {
     "duration": 0.044687,
     "end_time": "2023-10-28T03:01:43.972537",
     "exception": false,
     "start_time": "2023-10-28T03:01:43.927850",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Value of Scalar is: 7\n"
     ]
    }
   ],
   "source": [
    "# Printing the value of the scalar tensor using formatted string.\n",
    "print(f\"Value of Scalar is: {scalar.item()}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "103196e4",
   "metadata": {
    "papermill": {
     "duration": 0.039418,
     "end_time": "2023-10-28T03:01:44.050864",
     "exception": false,
     "start_time": "2023-10-28T03:01:44.011446",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "#### **1D Vector**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "8c579eeb",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:44.133061Z",
     "iopub.status.busy": "2023-10-28T03:01:44.132112Z",
     "iopub.status.idle": "2023-10-28T03:01:44.140414Z",
     "shell.execute_reply": "2023-10-28T03:01:44.139506Z"
    },
    "id": "H-wCNG9nM9EO",
    "outputId": "0873b72c-45ac-497b-cddc-4f5c145610a0",
    "papermill": {
     "duration": 0.050935,
     "end_time": "2023-10-28T03:01:44.142480",
     "exception": false,
     "start_time": "2023-10-28T03:01:44.091545",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tensor([7, 7])"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Creating a 1-dimensional tensor (vector) with two elements, both being the value 7.\n",
    "vector = torch.tensor([7, 7])\n",
    "vector"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "6aed3f73",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:44.224521Z",
     "iopub.status.busy": "2023-10-28T03:01:44.224160Z",
     "iopub.status.idle": "2023-10-28T03:01:44.229051Z",
     "shell.execute_reply": "2023-10-28T03:01:44.228087Z"
    },
    "id": "KZBGfKcCNG5s",
    "outputId": "dbcee822-06ca-4862-b866-0582cac7ce78",
    "papermill": {
     "duration": 0.048364,
     "end_time": "2023-10-28T03:01:44.231134",
     "exception": false,
     "start_time": "2023-10-28T03:01:44.182770",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Domention of Vector is: 1\n"
     ]
    }
   ],
   "source": [
    "# Printing the number of dimensions of the scalar tensor. Expecting 0 because it's a scalar.\n",
    "print(f\"Domention of Vector is: {vector.ndim}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "03769a5d",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:44.311371Z",
     "iopub.status.busy": "2023-10-28T03:01:44.311011Z",
     "iopub.status.idle": "2023-10-28T03:01:44.316246Z",
     "shell.execute_reply": "2023-10-28T03:01:44.315240Z"
    },
    "id": "cJqqHWt2NTz0",
    "outputId": "1431bfb4-958f-46b3-f336-7741c93ba55b",
    "papermill": {
     "duration": 0.046807,
     "end_time": "2023-10-28T03:01:44.318208",
     "exception": false,
     "start_time": "2023-10-28T03:01:44.271401",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Shape of the Vector: torch.Size([2])\n"
     ]
    }
   ],
   "source": [
    "print(f\"Shape of the Vector: {vector.shape}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "72b8678d",
   "metadata": {
    "papermill": {
     "duration": 0.038556,
     "end_time": "2023-10-28T03:01:44.395063",
     "exception": false,
     "start_time": "2023-10-28T03:01:44.356507",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "#### **2D Tensor: Matrix**\n",
    "In this code, a 2-dimensional tensor (matrix) named MATRIX is created with the specified values in a 2x2 format. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "0e98cfeb",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:44.474972Z",
     "iopub.status.busy": "2023-10-28T03:01:44.474672Z",
     "iopub.status.idle": "2023-10-28T03:01:44.482678Z",
     "shell.execute_reply": "2023-10-28T03:01:44.481534Z"
    },
    "id": "CqLWvezQNYcr",
    "outputId": "65df4e5c-6a90-4be5-b544-cb3149f2d414",
    "papermill": {
     "duration": 0.050992,
     "end_time": "2023-10-28T03:01:44.485082",
     "exception": false,
     "start_time": "2023-10-28T03:01:44.434090",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "tensor([[ 7,  8],\n",
      "        [ 9, 10]])\n"
     ]
    }
   ],
   "source": [
    "# Creating a 2-dimensional tensor (matrix) with the specified values, and displaying it.\n",
    "MATRIX = torch.tensor([[7, 8],\n",
    "                       [9, 10]])\n",
    "print(MATRIX)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "b0767b4f",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:44.565434Z",
     "iopub.status.busy": "2023-10-28T03:01:44.565120Z",
     "iopub.status.idle": "2023-10-28T03:01:44.570079Z",
     "shell.execute_reply": "2023-10-28T03:01:44.568955Z"
    },
    "id": "6aXuxUwqNo-V",
    "outputId": "5f6e6f49-1653-4973-f18a-bfdd3fe493ca",
    "papermill": {
     "duration": 0.047775,
     "end_time": "2023-10-28T03:01:44.572053",
     "exception": false,
     "start_time": "2023-10-28T03:01:44.524278",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of Dimensions of MATRIX: 2\n"
     ]
    }
   ],
   "source": [
    "# Printing the number of dimensions of the MATRIX tensor. Since MATRIX is 2-dimensional, it will return 2.\n",
    "print(f\"Number of Dimensions of MATRIX: {MATRIX.ndim}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "6ba12a5c",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:44.652336Z",
     "iopub.status.busy": "2023-10-28T03:01:44.651660Z",
     "iopub.status.idle": "2023-10-28T03:01:44.656544Z",
     "shell.execute_reply": "2023-10-28T03:01:44.655505Z"
    },
    "id": "KDNQ8WTJNrne",
    "outputId": "e8a6f040-7aef-48ee-fa35-3d9709ed2d5b",
    "papermill": {
     "duration": 0.046073,
     "end_time": "2023-10-28T03:01:44.658627",
     "exception": false,
     "start_time": "2023-10-28T03:01:44.612554",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Shape of MATRIX: torch.Size([2, 2])\n"
     ]
    }
   ],
   "source": [
    "# Printing the shape of the MATRIX tensor. This will return the size of each dimension of the MATRIX.\n",
    "print(f\"Shape of MATRIX: {MATRIX.shape}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "7ee9e7a3",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:44.737354Z",
     "iopub.status.busy": "2023-10-28T03:01:44.737057Z",
     "iopub.status.idle": "2023-10-28T03:01:44.742092Z",
     "shell.execute_reply": "2023-10-28T03:01:44.741144Z"
    },
    "id": "gsnAnJjCNubr",
    "outputId": "4250929e-18fb-4fd5-a485-04d60bb8762c",
    "papermill": {
     "duration": 0.045876,
     "end_time": "2023-10-28T03:01:44.744068",
     "exception": false,
     "start_time": "2023-10-28T03:01:44.698192",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "tensor([ 9, 10])\n"
     ]
    }
   ],
   "source": [
    "# Accessing and printing the second row of the MATRIX tensor.\n",
    "# Indexing is zero-based.\n",
    "print(MATRIX[1])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "ac94448c",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:44.820873Z",
     "iopub.status.busy": "2023-10-28T03:01:44.820611Z",
     "iopub.status.idle": "2023-10-28T03:01:44.825733Z",
     "shell.execute_reply": "2023-10-28T03:01:44.824735Z"
    },
    "id": "DidangZmNxv1",
    "outputId": "3e5daa1a-83d1-417b-d14f-36ff4be3e25f",
    "papermill": {
     "duration": 0.045827,
     "end_time": "2023-10-28T03:01:44.827598",
     "exception": false,
     "start_time": "2023-10-28T03:01:44.781771",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "tensor([7, 8])\n"
     ]
    }
   ],
   "source": [
    "# Accessing and printing the first row of the MATRIX tensor.\n",
    "print(MATRIX[0])"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6ce0118a",
   "metadata": {
    "papermill": {
     "duration": 0.037324,
     "end_time": "2023-10-28T03:01:44.902496",
     "exception": false,
     "start_time": "2023-10-28T03:01:44.865172",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "#### **3D Tensor**\n",
    "In this code snippet, a 3-dimensional tensor named TENSOR is created with the specified values. Using the print statement will cleanly display the TENSOR tensor when you run this code, showcasing its three-dimensional structure:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "4ef97107",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:44.979151Z",
     "iopub.status.busy": "2023-10-28T03:01:44.978867Z",
     "iopub.status.idle": "2023-10-28T03:01:44.984773Z",
     "shell.execute_reply": "2023-10-28T03:01:44.983826Z"
    },
    "id": "Dwbbm-R9NzB_",
    "outputId": "5a30b5ce-b94e-4c83-8239-13d5394702f1",
    "papermill": {
     "duration": 0.046586,
     "end_time": "2023-10-28T03:01:44.986745",
     "exception": false,
     "start_time": "2023-10-28T03:01:44.940159",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "tensor([[[1, 2, 3],\n",
      "         [3, 4, 5],\n",
      "         [5, 6, 7]]])\n"
     ]
    }
   ],
   "source": [
    "# Creating a 3-dimensional tensor (TENSOR) with the specified values, and displaying it.\n",
    "TENSOR = torch.tensor([[[1, 2, 3],\n",
    "                        [3, 4, 5],\n",
    "                        [5, 6, 7]]])\n",
    "print(TENSOR)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "12dcfee6",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:45.065945Z",
     "iopub.status.busy": "2023-10-28T03:01:45.065657Z",
     "iopub.status.idle": "2023-10-28T03:01:45.070400Z",
     "shell.execute_reply": "2023-10-28T03:01:45.069489Z"
    },
    "id": "uj1AoYL3OLyH",
    "outputId": "018e7bad-95ba-4f9c-ae4c-b37a7a6c1c98",
    "papermill": {
     "duration": 0.047501,
     "end_time": "2023-10-28T03:01:45.072438",
     "exception": false,
     "start_time": "2023-10-28T03:01:45.024937",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Shape of TENSOR: torch.Size([1, 3, 3]), Number of Dimensions: 3\n"
     ]
    }
   ],
   "source": [
    "# Printing the shape and number of dimensions of the TENSOR. \n",
    "# The shape will return the size of each dimension, and ndim will return the number of dimensions.\n",
    "print(f\"Shape of TENSOR: {TENSOR.shape}, Number of Dimensions: {TENSOR.ndim}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "81e782ba",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:45.151959Z",
     "iopub.status.busy": "2023-10-28T03:01:45.151681Z",
     "iopub.status.idle": "2023-10-28T03:01:45.157307Z",
     "shell.execute_reply": "2023-10-28T03:01:45.156348Z"
    },
    "id": "fOaZYE4aOQkM",
    "outputId": "30d3c14d-b5ab-4d68-afc4-ac408c6e9848",
    "papermill": {
     "duration": 0.048194,
     "end_time": "2023-10-28T03:01:45.159344",
     "exception": false,
     "start_time": "2023-10-28T03:01:45.111150",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "tensor([[1, 2, 3],\n",
      "        [3, 4, 5],\n",
      "        [5, 6, 7]])\n"
     ]
    }
   ],
   "source": [
    "# Accessing and printing the first element of the TENSOR. Since TENSOR is a 3D tensor, this will return a 2D matrix.\n",
    "print(TENSOR[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "fb20e0e4",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:45.239249Z",
     "iopub.status.busy": "2023-10-28T03:01:45.238565Z",
     "iopub.status.idle": "2023-10-28T03:01:45.244714Z",
     "shell.execute_reply": "2023-10-28T03:01:45.243823Z"
    },
    "id": "yUnBeYyJOeMU",
    "outputId": "3a8050f4-7330-4b1d-f396-c8274ef239e6",
    "papermill": {
     "duration": 0.048335,
     "end_time": "2023-10-28T03:01:45.246645",
     "exception": false,
     "start_time": "2023-10-28T03:01:45.198310",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "tensor([[[1, 2, 3],\n",
      "         [3, 4, 5],\n",
      "         [5, 6, 7]],\n",
      "\n",
      "        [[1, 2, 3],\n",
      "         [3, 4, 5],\n",
      "         [5, 6, 7]]])\n"
     ]
    }
   ],
   "source": [
    "# 3D Tensor\n",
    "# Creating a 3-dimensional tensor (TENSOR_2) with two 3x3 matrices, and displaying it.\n",
    "TENSOR_2 = torch.tensor([[[1, 2, 3],\n",
    "                          [3, 4, 5],\n",
    "                          [5, 6, 7]],\n",
    "\n",
    "                         [[1, 2, 3],\n",
    "                          [3, 4, 5],\n",
    "                          [5, 6, 7]]\n",
    "                         ])\n",
    "\n",
    "print(TENSOR_2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "80781abb",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:45.325970Z",
     "iopub.status.busy": "2023-10-28T03:01:45.325454Z",
     "iopub.status.idle": "2023-10-28T03:01:45.330071Z",
     "shell.execute_reply": "2023-10-28T03:01:45.329297Z"
    },
    "id": "cHaGNCnGOlZb",
    "outputId": "00fcd6a5-3758-4b47-c8e9-d3e4bdbf3332",
    "papermill": {
     "duration": 0.045738,
     "end_time": "2023-10-28T03:01:45.332106",
     "exception": false,
     "start_time": "2023-10-28T03:01:45.286368",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of Dimensions of TENSOR_2: 3, Shape of TENSOR_2: torch.Size([2, 3, 3])\n"
     ]
    }
   ],
   "source": [
    "# Printing the number of dimensions and the shape of TENSOR_2.\n",
    "# .ndim returns the number of dimensions, and .shape returns the size of each dimension.\n",
    "print(f\"Number of Dimensions of TENSOR_2: {TENSOR_2.ndim}, Shape of TENSOR_2: {TENSOR_2.shape}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "7df6f42e",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:45.409669Z",
     "iopub.status.busy": "2023-10-28T03:01:45.409406Z",
     "iopub.status.idle": "2023-10-28T03:01:45.415140Z",
     "shell.execute_reply": "2023-10-28T03:01:45.414288Z"
    },
    "id": "ntgeCd8WPKfU",
    "outputId": "7b132d38-4975-472e-938c-af6085370cc8",
    "papermill": {
     "duration": 0.046881,
     "end_time": "2023-10-28T03:01:45.417370",
     "exception": false,
     "start_time": "2023-10-28T03:01:45.370489",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of Dimensions of TENSOR_3 is: 3, Shape of TENSOR_3: torch.Size([3, 3, 3])\n"
     ]
    }
   ],
   "source": [
    "# 3D Tensor\n",
    "# Creating a 3-dimensional tensor (TENSOR_3) with three 3x3 matrices, and displaying its number of dimensions and shape.\n",
    "TENSOR_3 = torch.tensor([[[1, 2, 3],\n",
    "                          [3, 4, 5],\n",
    "                          [5, 6, 7]],\n",
    "\n",
    "                         [[1, 2, 3],\n",
    "                          [3, 4, 5],\n",
    "                          [5, 6, 7]],\n",
    "\n",
    "                         [[1, 2, 3],\n",
    "                          [3, 4, 5],\n",
    "                          [5, 6, 7]]\n",
    "                         ])\n",
    "\n",
    "# Printing the number of dimensions and the shape of TENSOR_3.\n",
    "print(f\"Number of Dimensions of TENSOR_3 is: {TENSOR_3.ndim}, Shape of TENSOR_3: {TENSOR_3.shape}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d7acdba5",
   "metadata": {
    "id": "P94566UIPZk8",
    "papermill": {
     "duration": 0.039028,
     "end_time": "2023-10-28T03:01:45.494346",
     "exception": false,
     "start_time": "2023-10-28T03:01:45.455318",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "## **Creating Random Tensors**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "d65fb96b",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:45.574931Z",
     "iopub.status.busy": "2023-10-28T03:01:45.574563Z",
     "iopub.status.idle": "2023-10-28T03:01:45.629222Z",
     "shell.execute_reply": "2023-10-28T03:01:45.628126Z"
    },
    "id": "Xhqso6WxPg3X",
    "outputId": "5e19117c-d5e6-4f1c-f5df-402188337b02",
    "papermill": {
     "duration": 0.097813,
     "end_time": "2023-10-28T03:01:45.631451",
     "exception": false,
     "start_time": "2023-10-28T03:01:45.533638",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "tensor([[0.3651, 0.0635, 0.8533, 0.8187],\n",
      "        [0.1126, 0.6958, 0.8192, 0.6821],\n",
      "        [0.9119, 0.8845, 0.3161, 0.4256]])\n"
     ]
    }
   ],
   "source": [
    "# Creating a random tensor of shape (3, 4) with values uniformly distributed between 0 and 1.\n",
    "random_tensor = torch.rand(3, 4)\n",
    "print(random_tensor)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "4771f7c1",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:45.718683Z",
     "iopub.status.busy": "2023-10-28T03:01:45.718346Z",
     "iopub.status.idle": "2023-10-28T03:01:45.723086Z",
     "shell.execute_reply": "2023-10-28T03:01:45.722187Z"
    },
    "id": "mvnTVgo1T7Rc",
    "outputId": "2c15dcfe-441c-4363-b276-9b1f31c9ff86",
    "papermill": {
     "duration": 0.050426,
     "end_time": "2023-10-28T03:01:45.725090",
     "exception": false,
     "start_time": "2023-10-28T03:01:45.674664",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of Dimensions of random_tensor: 2\n"
     ]
    }
   ],
   "source": [
    "# Printing the number of dimensions of random_tensor using the .ndim attribute.\n",
    "print(f\"Number of Dimensions of random_tensor: {random_tensor.ndim}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "efd7abad",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:45.809807Z",
     "iopub.status.busy": "2023-10-28T03:01:45.809545Z",
     "iopub.status.idle": "2023-10-28T03:01:45.813880Z",
     "shell.execute_reply": "2023-10-28T03:01:45.813038Z"
    },
    "id": "d2MnyFbpT_Ad",
    "outputId": "cde50133-e95b-412f-da3d-992cd0b4f5aa",
    "papermill": {
     "duration": 0.046759,
     "end_time": "2023-10-28T03:01:45.816002",
     "exception": false,
     "start_time": "2023-10-28T03:01:45.769243",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Shape of random_tensor: torch.Size([3, 4])\n"
     ]
    }
   ],
   "source": [
    "# Printing the shape of random_tensor using the .shape attribute.\n",
    "# This will return a tuple representing the size of each dimension.\n",
    "print(f\"Shape of random_tensor: {random_tensor.shape}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "487f5061",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:45.893995Z",
     "iopub.status.busy": "2023-10-28T03:01:45.893294Z",
     "iopub.status.idle": "2023-10-28T03:01:45.899173Z",
     "shell.execute_reply": "2023-10-28T03:01:45.898154Z"
    },
    "id": "UAdKHLhPUBVh",
    "outputId": "712073a3-83b2-4230-df23-28940ebc7a19",
    "papermill": {
     "duration": 0.046893,
     "end_time": "2023-10-28T03:01:45.901198",
     "exception": false,
     "start_time": "2023-10-28T03:01:45.854305",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "tensor([[[0.2135, 0.5271, 0.7504, 0.3874],\n",
      "         [0.3376, 0.6081, 0.3399, 0.4245],\n",
      "         [0.7120, 0.1783, 0.0025, 0.2874],\n",
      "         [0.4554, 0.0897, 0.0881, 0.4283],\n",
      "         [0.6020, 0.3639, 0.0040, 0.6427],\n",
      "         [0.7548, 0.0878, 0.9973, 0.5440]]])\n"
     ]
    }
   ],
   "source": [
    "# Creating a 3D random tensor of shape (1, 6, 4) with values uniformly distributed between 0 and 1.\n",
    "random_tensor = torch.rand(1, 6, 4)\n",
    "print(random_tensor)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "c2d6af91",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:45.981099Z",
     "iopub.status.busy": "2023-10-28T03:01:45.980836Z",
     "iopub.status.idle": "2023-10-28T03:01:45.987626Z",
     "shell.execute_reply": "2023-10-28T03:01:45.986539Z"
    },
    "id": "5NVV5hE1UZ1X",
    "outputId": "fa567d2f-2962-4654-8944-e5fac4a5f8dd",
    "papermill": {
     "duration": 0.049624,
     "end_time": "2023-10-28T03:01:45.989611",
     "exception": false,
     "start_time": "2023-10-28T03:01:45.939987",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Shape of random_image_size_tensor: torch.Size([224, 224, 3]), Number of Dimensions: 3\n"
     ]
    }
   ],
   "source": [
    "# Create a random tensor resembling an image size\n",
    "# Generating a random tensor of shape (224, 224, 3), symbolizing an image of 224x224 pixels with 3 color channels (RGB).\n",
    "random_image_size_tensor = torch.rand(size=(224, 224, 3))\n",
    "\n",
    "# Printing the shape and number of dimensions of the generated tensor.\n",
    "print(f\"Shape of random_image_size_tensor: {random_image_size_tensor.shape}, Number of Dimensions: {random_image_size_tensor.ndim}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7891b25c",
   "metadata": {
    "id": "zRb7GVNqVm2d",
    "papermill": {
     "duration": 0.040929,
     "end_time": "2023-10-28T03:01:46.069793",
     "exception": false,
     "start_time": "2023-10-28T03:01:46.028864",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "## **Zeros and Ones**\n",
    "In PyTorch, you can easily create tensors filled exclusively with zeros or ones. These types of tensors can be particularly useful for various scenarios, such as initializing weights in neural networks, among other use cases."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "229cad1e",
   "metadata": {
    "papermill": {
     "duration": 0.038217,
     "end_time": "2023-10-28T03:01:46.146711",
     "exception": false,
     "start_time": "2023-10-28T03:01:46.108494",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "#### **Creating a Zero Tensor**\n",
    "You can use the torch.zeros() function to create a tensor filled with zeros. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "c35e6d20",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:46.229145Z",
     "iopub.status.busy": "2023-10-28T03:01:46.228848Z",
     "iopub.status.idle": "2023-10-28T03:01:46.236221Z",
     "shell.execute_reply": "2023-10-28T03:01:46.235218Z"
    },
    "id": "YiRiJPF-VpXP",
    "outputId": "623da6c1-93b9-4d34-c753-f79814fa10b5",
    "papermill": {
     "duration": 0.051917,
     "end_time": "2023-10-28T03:01:46.238161",
     "exception": false,
     "start_time": "2023-10-28T03:01:46.186244",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Zero Tensor:\n",
      "tensor([[0., 0., 0.],\n",
      "        [0., 0., 0.],\n",
      "        [0., 0., 0.]])\n"
     ]
    }
   ],
   "source": [
    "# Creating a tensor of shape (3, 3) filled with zeros\n",
    "zero_tensor = torch.zeros(3, 3)\n",
    "print(\"Zero Tensor:\")\n",
    "print(zero_tensor)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ed7af709",
   "metadata": {
    "papermill": {
     "duration": 0.041239,
     "end_time": "2023-10-28T03:01:46.321121",
     "exception": false,
     "start_time": "2023-10-28T03:01:46.279882",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "#### **Creating a One Tensor**\n",
    "Likewise, the torch.ones() function allows you to create a tensor populated with ones. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "bc2f0e73",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:46.401994Z",
     "iopub.status.busy": "2023-10-28T03:01:46.401340Z",
     "iopub.status.idle": "2023-10-28T03:01:46.408211Z",
     "shell.execute_reply": "2023-10-28T03:01:46.407287Z"
    },
    "id": "ZAmK47xXVzAn",
    "outputId": "feed3f13-7db4-473b-821c-d47518f6c86f",
    "papermill": {
     "duration": 0.049568,
     "end_time": "2023-10-28T03:01:46.410153",
     "exception": false,
     "start_time": "2023-10-28T03:01:46.360585",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "One Tensor:\n",
      "tensor([[1., 1., 1.],\n",
      "        [1., 1., 1.],\n",
      "        [1., 1., 1.]])\n"
     ]
    }
   ],
   "source": [
    "# Creating a tensor of shape (3, 3) filled with ones\n",
    "one_tensor = torch.ones(3, 3)\n",
    "print(\"\\nOne Tensor:\")\n",
    "print(one_tensor)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "50ae268a",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:46.491303Z",
     "iopub.status.busy": "2023-10-28T03:01:46.490791Z",
     "iopub.status.idle": "2023-10-28T03:01:46.495946Z",
     "shell.execute_reply": "2023-10-28T03:01:46.495008Z"
    },
    "id": "9vtFO7b5V31M",
    "outputId": "f4cfc5f4-85ba-4426-9861-0d62f404803c",
    "papermill": {
     "duration": 0.048277,
     "end_time": "2023-10-28T03:01:46.497999",
     "exception": false,
     "start_time": "2023-10-28T03:01:46.449722",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Data type of zero_tensor: torch.float32\n",
      "Data type of one_tensor: torch.float32\n"
     ]
    }
   ],
   "source": [
    "# Displaying the data types of zero_tensor and one_tensor using the .dtype attribute.\n",
    "# It returns the data type of the tensor's elements.\n",
    "print(f\"Data type of zero_tensor: {zero_tensor.dtype}\")\n",
    "print(f\"Data type of one_tensor: {one_tensor.dtype}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6f81c9ab",
   "metadata": {
    "id": "ism2q17AV54A",
    "papermill": {
     "duration": 0.039817,
     "end_time": "2023-10-28T03:01:46.579468",
     "exception": false,
     "start_time": "2023-10-28T03:01:46.539651",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "## **Arange of Tensors and Tensor-like**\n",
    "The torch.arange() function is used to create a 1-dimensional tensor with a sequence of numbers within a specified range. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "4e96759d",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:46.660245Z",
     "iopub.status.busy": "2023-10-28T03:01:46.659931Z",
     "iopub.status.idle": "2023-10-28T03:01:46.667823Z",
     "shell.execute_reply": "2023-10-28T03:01:46.666937Z"
    },
    "id": "j62QoSg6WP4q",
    "outputId": "4368008d-36ac-497a-d013-0a19f717746b",
    "papermill": {
     "duration": 0.05002,
     "end_time": "2023-10-28T03:01:46.669772",
     "exception": false,
     "start_time": "2023-10-28T03:01:46.619752",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Tensor created with torch.arange(0, 10):\n",
      "tensor([0, 1, 2, 3, 4, 5, 6, 7, 8, 9])\n"
     ]
    }
   ],
   "source": [
    "# Creating a 1D tensor with a sequence of integers from 0 to 9.\n",
    "arange_tensor = torch.arange(0, 10)\n",
    "\n",
    "# Printing the created tensor.\n",
    "print(\"Tensor created with torch.arange(0, 10):\")\n",
    "print(arange_tensor)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "790077bd",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:46.750367Z",
     "iopub.status.busy": "2023-10-28T03:01:46.750079Z",
     "iopub.status.idle": "2023-10-28T03:01:46.755867Z",
     "shell.execute_reply": "2023-10-28T03:01:46.755008Z"
    },
    "id": "8RJq3S3iWT7A",
    "outputId": "24e31c6d-d2e8-47de-8eb7-2d017aa85b21",
    "papermill": {
     "duration": 0.048255,
     "end_time": "2023-10-28T03:01:46.757922",
     "exception": false,
     "start_time": "2023-10-28T03:01:46.709667",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Tensor created with torch.arange(start=1, end=1000, step=77):\n",
      "tensor([  1,  78, 155, 232, 309, 386, 463, 540, 617, 694, 771, 848, 925])\n"
     ]
    }
   ],
   "source": [
    "# Using torch.arange() with specified start, end, and step values.\n",
    "# Creating a tensor with values starting from 1, ending before 1000, incrementing by 77.\n",
    "one_to_thousand = torch.arange(start=1, end=1000, step=77)\n",
    "\n",
    "# Printing the created tensor.\n",
    "print(\"Tensor created with torch.arange(start=1, end=1000, step=77):\")\n",
    "print(one_to_thousand)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "c00b32bd",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:46.839722Z",
     "iopub.status.busy": "2023-10-28T03:01:46.838850Z",
     "iopub.status.idle": "2023-10-28T03:01:46.844664Z",
     "shell.execute_reply": "2023-10-28T03:01:46.843801Z"
    },
    "id": "WTcrxrrJXGun",
    "outputId": "9a4bf4bb-e7cd-44b4-e5a9-0979a47dad6d",
    "papermill": {
     "duration": 0.04919,
     "end_time": "2023-10-28T03:01:46.846583",
     "exception": false,
     "start_time": "2023-10-28T03:01:46.797393",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "torch.Size([13])"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "one_to_thousand.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6db59870",
   "metadata": {
    "papermill": {
     "duration": 0.039306,
     "end_time": "2023-10-28T03:01:46.925093",
     "exception": false,
     "start_time": "2023-10-28T03:01:46.885787",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "#### **Tensor-Like**\n",
    "torch.zeros_like() is a practical function used in various scenarios, primarily when you need a new tensor with the same shape and properties (e.g., data type, device) as an existing tensor but want to initialize it with zeros. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "1d7c780d",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:47.004693Z",
     "iopub.status.busy": "2023-10-28T03:01:47.004383Z",
     "iopub.status.idle": "2023-10-28T03:01:47.010547Z",
     "shell.execute_reply": "2023-10-28T03:01:47.009647Z"
    },
    "id": "myLdQGs1W-VB",
    "outputId": "0973f94e-cc7c-458b-d18f-ba065161a53e",
    "papermill": {
     "duration": 0.048584,
     "end_time": "2023-10-28T03:01:47.012538",
     "exception": false,
     "start_time": "2023-10-28T03:01:46.963954",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Tensor ten_zeros, with the same shape and properties as one_to_ten but filled with zeros:\n",
      "tensor([0, 0, 0, 0, 0, 0, 0, 0, 0, 0])\n"
     ]
    }
   ],
   "source": [
    "one_to_ten = torch.arange(1, 11)  # Creating a tensor with values from 1 to 10\n",
    "\n",
    "# Creating a tensor named ten_zeros with the same shape and properties as one_to_ten, but filled with zeros.\n",
    "ten_zeros = torch.zeros_like(input=one_to_ten)\n",
    "\n",
    "# Printing the ten_zeros tensor\n",
    "print(\"Tensor ten_zeros, with the same shape and properties as one_to_ten but filled with zeros:\")\n",
    "print(ten_zeros)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6dd1406d",
   "metadata": {
    "id": "t92ZQdAzXPLI",
    "papermill": {
     "duration": 0.040409,
     "end_time": "2023-10-28T03:01:47.094238",
     "exception": false,
     "start_time": "2023-10-28T03:01:47.053829",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "## **Tensor Data Types**"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6cb74a8c",
   "metadata": {
    "papermill": {
     "duration": 0.040472,
     "end_time": "2023-10-28T03:01:47.174422",
     "exception": false,
     "start_time": "2023-10-28T03:01:47.133950",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "#### **Float 16**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "id": "18dbca20",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:47.256917Z",
     "iopub.status.busy": "2023-10-28T03:01:47.256568Z",
     "iopub.status.idle": "2023-10-28T03:01:47.266766Z",
     "shell.execute_reply": "2023-10-28T03:01:47.265786Z"
    },
    "papermill": {
     "duration": 0.054451,
     "end_time": "2023-10-28T03:01:47.269277",
     "exception": false,
     "start_time": "2023-10-28T03:01:47.214826",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Tensor: tensor([3., 6., 9.], dtype=torch.float16), Data Type: torch.float16\n"
     ]
    }
   ],
   "source": [
    "# Creating a tensor with floating-point values and specifying the data type as torch.float16.\n",
    "float_16_tensor = torch.tensor([3.0, 6.0, 9.0], dtype=torch.float16)\n",
    "\n",
    "# Printing the tensor and its data type.\n",
    "print(f\"Tensor: {float_16_tensor}, Data Type: {float_16_tensor.dtype}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b32e426f",
   "metadata": {
    "papermill": {
     "duration": 0.039316,
     "end_time": "2023-10-28T03:01:47.349024",
     "exception": false,
     "start_time": "2023-10-28T03:01:47.309708",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "#### **Float 32**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "id": "26f5402d",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:47.431541Z",
     "iopub.status.busy": "2023-10-28T03:01:47.430858Z",
     "iopub.status.idle": "2023-10-28T03:01:47.437027Z",
     "shell.execute_reply": "2023-10-28T03:01:47.435986Z"
    },
    "id": "Eaj8hiWIXq2M",
    "outputId": "b3ada270-0d87-4ab0-8bbd-7b175746edd6",
    "papermill": {
     "duration": 0.049624,
     "end_time": "2023-10-28T03:01:47.438995",
     "exception": false,
     "start_time": "2023-10-28T03:01:47.389371",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Tensor: tensor([3., 6., 9.]), Data Type: torch.float32\n"
     ]
    }
   ],
   "source": [
    "# Creating a tensor with floating-point values and letting PyTorch infer the data type.\n",
    "float_32_tensor = torch.tensor([3.0, 6.0, 9.0], dtype=None)\n",
    "\n",
    "# Printing the tensor and its data type.\n",
    "print(f\"Tensor: {float_32_tensor}, Data Type: {float_32_tensor.dtype}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "id": "f82963e7",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:47.523421Z",
     "iopub.status.busy": "2023-10-28T03:01:47.523105Z",
     "iopub.status.idle": "2023-10-28T03:01:47.530008Z",
     "shell.execute_reply": "2023-10-28T03:01:47.529038Z"
    },
    "id": "FlJ61GoiX-zE",
    "outputId": "434c94c2-f267-4cfe-d6a7-43c0d0e8e494",
    "papermill": {
     "duration": 0.053001,
     "end_time": "2023-10-28T03:01:47.532176",
     "exception": false,
     "start_time": "2023-10-28T03:01:47.479175",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Tensor: tensor([3., 6., 9.]), Data Type: torch.float32\n"
     ]
    }
   ],
   "source": [
    "# Creating a tensor with additional specifications: dtype, device, and requires_grad.\n",
    "float_32_tensor = torch.tensor([3.0, 6.0, 9.0],  # Data values\n",
    "                               dtype=None,        # Data Type: Not explicitly set, will be inferred\n",
    "                               device=None,       # Device: Not explicitly set, will use the default device (usually CPU)\n",
    "                               requires_grad=False)  # Track Gradients: False\n",
    "\n",
    "# Printing the tensor and its data type.\n",
    "print(f\"Tensor: {float_32_tensor}, Data Type: {float_32_tensor.dtype}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "id": "516a811e",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:47.617721Z",
     "iopub.status.busy": "2023-10-28T03:01:47.617435Z",
     "iopub.status.idle": "2023-10-28T03:01:47.623451Z",
     "shell.execute_reply": "2023-10-28T03:01:47.622404Z"
    },
    "id": "btGExBmlZ4Xk",
    "outputId": "257b6f6d-909d-4196-b248-d745cce65b8a",
    "papermill": {
     "duration": 0.051549,
     "end_time": "2023-10-28T03:01:47.625651",
     "exception": false,
     "start_time": "2023-10-28T03:01:47.574102",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Result Tensor: tensor([ 9., 36., 81.]), Data Type: torch.float32\n"
     ]
    }
   ],
   "source": [
    "# Element-wise multiplication between two tensors: float_16_tensor and float_32_tensor.\n",
    "nn = float_16_tensor * float_32_tensor\n",
    "\n",
    "# Printing the result tensor and its data type.\n",
    "print(f\"Result Tensor: {nn}, Data Type: {nn.dtype}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "id": "224d8b20",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:47.710106Z",
     "iopub.status.busy": "2023-10-28T03:01:47.709828Z",
     "iopub.status.idle": "2023-10-28T03:01:47.715291Z",
     "shell.execute_reply": "2023-10-28T03:01:47.714427Z"
    },
    "id": "zPMod49uaYAJ",
    "outputId": "2cba4580-39ed-4a4a-f571-ddc46182606e",
    "papermill": {
     "duration": 0.048174,
     "end_time": "2023-10-28T03:01:47.717288",
     "exception": false,
     "start_time": "2023-10-28T03:01:47.669114",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "tensor([3, 4, 2], dtype=torch.int32)\n"
     ]
    }
   ],
   "source": [
    "# Creating a tensor with integer values and specifying the data type as torch.int32.\n",
    "int_32_tensor = torch.tensor([3, 4, 2], dtype=torch.int32)\n",
    "\n",
    "# Printing the created tensor.\n",
    "print(int_32_tensor)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "id": "1b83a182",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:47.798813Z",
     "iopub.status.busy": "2023-10-28T03:01:47.798559Z",
     "iopub.status.idle": "2023-10-28T03:01:47.806605Z",
     "shell.execute_reply": "2023-10-28T03:01:47.805701Z"
    },
    "id": "ibrEoYhiaoLr",
    "outputId": "0a205e02-1baf-4294-b6d7-7d4907616909",
    "papermill": {
     "duration": 0.051101,
     "end_time": "2023-10-28T03:01:47.808547",
     "exception": false,
     "start_time": "2023-10-28T03:01:47.757446",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tensor([ 9., 24., 18.], dtype=torch.float16)"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "float_16_tensor*int_32_tensor"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3b4e7a95",
   "metadata": {
    "id": "4uC9teP_bT_K",
    "papermill": {
     "duration": 0.041076,
     "end_time": "2023-10-28T03:01:47.891614",
     "exception": false,
     "start_time": "2023-10-28T03:01:47.850538",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "## **Manipulation of Tensors**"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7c87ed25",
   "metadata": {
    "id": "deQ7qc3YdjEM",
    "papermill": {
     "duration": 0.041124,
     "end_time": "2023-10-28T03:01:47.975007",
     "exception": false,
     "start_time": "2023-10-28T03:01:47.933883",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "In PyTorch, tensors can be manipulated in various ways to perform mathematical operations, reshape their dimensions, and more. This section explores some common tensor operations:"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6d95455f",
   "metadata": {
    "papermill": {
     "duration": 0.042303,
     "end_time": "2023-10-28T03:01:48.059246",
     "exception": false,
     "start_time": "2023-10-28T03:01:48.016943",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "#### **1. Tensor Arithmetic Operations**\n",
    "You can perform basic arithmetic operations on tensors, such as addition, subtraction, multiplication, and division. For example:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "id": "8ecba6d8",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:48.145514Z",
     "iopub.status.busy": "2023-10-28T03:01:48.144681Z",
     "iopub.status.idle": "2023-10-28T03:01:48.154039Z",
     "shell.execute_reply": "2023-10-28T03:01:48.153107Z"
    },
    "papermill": {
     "duration": 0.054635,
     "end_time": "2023-10-28T03:01:48.156086",
     "exception": false,
     "start_time": "2023-10-28T03:01:48.101451",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tensor([5, 7, 9])"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Creating two tensors\n",
    "tensor1 = torch.tensor([1, 2, 3])\n",
    "tensor2 = torch.tensor([4, 5, 6])\n",
    "\n",
    "# Addition\n",
    "result_addition = tensor1 + tensor2\n",
    "result_addition"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "id": "e54246f5",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:48.238493Z",
     "iopub.status.busy": "2023-10-28T03:01:48.238183Z",
     "iopub.status.idle": "2023-10-28T03:01:48.245185Z",
     "shell.execute_reply": "2023-10-28T03:01:48.244309Z"
    },
    "papermill": {
     "duration": 0.050985,
     "end_time": "2023-10-28T03:01:48.247091",
     "exception": false,
     "start_time": "2023-10-28T03:01:48.196106",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tensor([-3, -3, -3])"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Subtraction\n",
    "result_subtraction = tensor1 - tensor2\n",
    "result_subtraction"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "id": "02f7f3b3",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:48.331676Z",
     "iopub.status.busy": "2023-10-28T03:01:48.330867Z",
     "iopub.status.idle": "2023-10-28T03:01:48.338583Z",
     "shell.execute_reply": "2023-10-28T03:01:48.337875Z"
    },
    "papermill": {
     "duration": 0.050946,
     "end_time": "2023-10-28T03:01:48.340370",
     "exception": false,
     "start_time": "2023-10-28T03:01:48.289424",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tensor([ 4, 10, 18])"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Multiplication\n",
    "result_multiplication = tensor1 * tensor2\n",
    "result_multiplication"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "id": "9a3a7e6c",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:48.425871Z",
     "iopub.status.busy": "2023-10-28T03:01:48.425543Z",
     "iopub.status.idle": "2023-10-28T03:01:48.432331Z",
     "shell.execute_reply": "2023-10-28T03:01:48.431303Z"
    },
    "papermill": {
     "duration": 0.051684,
     "end_time": "2023-10-28T03:01:48.434485",
     "exception": false,
     "start_time": "2023-10-28T03:01:48.382801",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tensor([0.2500, 0.4000, 0.5000])"
      ]
     },
     "execution_count": 40,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Division\n",
    "result_division = tensor1 / tensor2\n",
    "result_division"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "df90ad97",
   "metadata": {
    "papermill": {
     "duration": 0.040848,
     "end_time": "2023-10-28T03:01:48.516083",
     "exception": false,
     "start_time": "2023-10-28T03:01:48.475235",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "#### **2. Reshaping Tensors**\n",
    "You can change the shape of a tensor using the .reshape() method or the .view() method. For example:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "id": "7a31f41d",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:48.602006Z",
     "iopub.status.busy": "2023-10-28T03:01:48.601172Z",
     "iopub.status.idle": "2023-10-28T03:01:48.606576Z",
     "shell.execute_reply": "2023-10-28T03:01:48.605568Z"
    },
    "papermill": {
     "duration": 0.051315,
     "end_time": "2023-10-28T03:01:48.608633",
     "exception": false,
     "start_time": "2023-10-28T03:01:48.557318",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "# Reshaping a tensor to a different shape\n",
    "original_tensor = torch.tensor([[1, 2], [3, 4], [5, 6]])\n",
    "reshaped_tensor = original_tensor.reshape(2, 3)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6e2ccab0",
   "metadata": {
    "papermill": {
     "duration": 0.043346,
     "end_time": "2023-10-28T03:01:48.697813",
     "exception": false,
     "start_time": "2023-10-28T03:01:48.654467",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "#### **3. Reduction Operations**\n",
    "Reduction operations allow you to perform computations across the elements of a tensor. Common reduction operations include sum, mean, minimum, and maximum:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "id": "d459934c",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:48.796209Z",
     "iopub.status.busy": "2023-10-28T03:01:48.795562Z",
     "iopub.status.idle": "2023-10-28T03:01:48.809153Z",
     "shell.execute_reply": "2023-10-28T03:01:48.808151Z"
    },
    "papermill": {
     "duration": 0.067026,
     "end_time": "2023-10-28T03:01:48.811328",
     "exception": false,
     "start_time": "2023-10-28T03:01:48.744302",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Sum: tensor(15.)\n",
      "Mean: tensor(3.)\n",
      "Min: tensor(1.)\n",
      "Max: tensor(5.)\n"
     ]
    }
   ],
   "source": [
    "tensor = torch.tensor([1, 2, 3, 4, 5], dtype=torch.float32)  # Convert to float32\n",
    "sum_result = torch.sum(tensor)\n",
    "mean_result = torch.mean(tensor)\n",
    "min_result = torch.min(tensor)\n",
    "max_result = torch.max(tensor)\n",
    "\n",
    "print(\"Sum:\", sum_result)\n",
    "print(\"Mean:\", mean_result)\n",
    "print(\"Min:\", min_result)\n",
    "print(\"Max:\", max_result)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "edd5151a",
   "metadata": {
    "papermill": {
     "duration": 0.042486,
     "end_time": "2023-10-28T03:01:48.898235",
     "exception": false,
     "start_time": "2023-10-28T03:01:48.855749",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "#### **4. Element-wise Functions**\n",
    "PyTorch provides various element-wise functions, such as torch.sin(), torch.cos(), torch.exp(), and more, to apply mathematical functions to each element of a tensor:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "id": "24b534da",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:48.982809Z",
     "iopub.status.busy": "2023-10-28T03:01:48.982495Z",
     "iopub.status.idle": "2023-10-28T03:01:49.032102Z",
     "shell.execute_reply": "2023-10-28T03:01:49.031118Z"
    },
    "papermill": {
     "duration": 0.094154,
     "end_time": "2023-10-28T03:01:49.034285",
     "exception": false,
     "start_time": "2023-10-28T03:01:48.940131",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "# Element-wise functions\n",
    "tensor = torch.tensor([0.0, 1.0, 2.0])\n",
    "sin_result = torch.sin(tensor)\n",
    "exp_result = torch.exp(tensor)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2ddc940b",
   "metadata": {
    "papermill": {
     "duration": 0.040923,
     "end_time": "2023-10-28T03:01:49.118136",
     "exception": false,
     "start_time": "2023-10-28T03:01:49.077213",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "#### **5. Indexing and Slicing**\n",
    "You can access specific elements or slices of a tensor using indexing and slicing operations:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "id": "0f0c5b9f",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:49.202179Z",
     "iopub.status.busy": "2023-10-28T03:01:49.201885Z",
     "iopub.status.idle": "2023-10-28T03:01:49.210379Z",
     "shell.execute_reply": "2023-10-28T03:01:49.209486Z"
    },
    "papermill": {
     "duration": 0.05234,
     "end_time": "2023-10-28T03:01:49.212401",
     "exception": false,
     "start_time": "2023-10-28T03:01:49.160061",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "# Indexing and slicing\n",
    "tensor = torch.tensor([1, 2, 3, 4, 5])\n",
    "element_at_index_2 = tensor[2]\n",
    "slice = tensor[1:4]"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6b12894c",
   "metadata": {
    "id": "oOxeo3rmduvK",
    "papermill": {
     "duration": 0.042381,
     "end_time": "2023-10-28T03:01:49.297012",
     "exception": false,
     "start_time": "2023-10-28T03:01:49.254631",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "\n",
    "PyTorch also has a bunch of built-in functions like torch.mul() (short for multiplication) and torch.add() to perform basic operations."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "04b0b713",
   "metadata": {
    "id": "AuuScX1Ed5N8",
    "papermill": {
     "duration": 0.040466,
     "end_time": "2023-10-28T03:01:49.379027",
     "exception": false,
     "start_time": "2023-10-28T03:01:49.338561",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "#### **6. Matrix Multiplication**\n",
    "Matrix multiplication, also known as matrix product or dot product, is a fundamental operation in linear algebra. In PyTorch, you can perform matrix multiplication efficiently using various methods."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "id": "5489dee7",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:49.468615Z",
     "iopub.status.busy": "2023-10-28T03:01:49.468228Z",
     "iopub.status.idle": "2023-10-28T03:01:49.475568Z",
     "shell.execute_reply": "2023-10-28T03:01:49.474166Z"
    },
    "id": "vsstfW4UfHDX",
    "outputId": "4364bd83-c899-429d-d2dd-63b3ab6539a4",
    "papermill": {
     "duration": 0.056234,
     "end_time": "2023-10-28T03:01:49.477783",
     "exception": false,
     "start_time": "2023-10-28T03:01:49.421549",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Tensor: tensor([1, 2, 3])\n",
      "Element-wise Multiplication Result: tensor([1, 4, 9])\n"
     ]
    }
   ],
   "source": [
    "# Creating a tensor\n",
    "tensor = torch.tensor([1, 2, 3])\n",
    "\n",
    "# Element-wise multiplication: Each element multiplies its equivalent (index 0->0, 1->1, 2->2)\n",
    "result = tensor * tensor\n",
    "\n",
    "# Printing the tensor and its element-wise multiplication\n",
    "print(\"Tensor:\", tensor)\n",
    "print(\"Element-wise Multiplication Result:\", result)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "id": "086db3f4",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:49.564079Z",
     "iopub.status.busy": "2023-10-28T03:01:49.563757Z",
     "iopub.status.idle": "2023-10-28T03:01:49.575041Z",
     "shell.execute_reply": "2023-10-28T03:01:49.574131Z"
    },
    "id": "WJ9YeeSRfSpS",
    "outputId": "b8ac4a25-2142-49d8-86b2-109e7081fa86",
    "papermill": {
     "duration": 0.055995,
     "end_time": "2023-10-28T03:01:49.576985",
     "exception": false,
     "start_time": "2023-10-28T03:01:49.520990",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tensor([[19, 22],\n",
       "        [43, 50]])"
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Create two matrices\n",
    "matrix1 = torch.tensor([[1, 2], [3, 4]])\n",
    "matrix2 = torch.tensor([[5, 6], [7, 8]])\n",
    "\n",
    "# Perform matrix multiplication using torch.matmul()\n",
    "result_matmul = torch.matmul(matrix1, matrix2)\n",
    "result_matmul"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "afff7712",
   "metadata": {
    "papermill": {
     "duration": 0.041077,
     "end_time": "2023-10-28T03:01:49.658856",
     "exception": false,
     "start_time": "2023-10-28T03:01:49.617779",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "*You can use the torch.matmul() function to perform matrix multiplication. It takes two tensors as arguments and returns their matrix product.*"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "id": "94142781",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:49.741090Z",
     "iopub.status.busy": "2023-10-28T03:01:49.740837Z",
     "iopub.status.idle": "2023-10-28T03:01:49.747294Z",
     "shell.execute_reply": "2023-10-28T03:01:49.746430Z"
    },
    "id": "CgJHx1LNfVv4",
    "outputId": "a32717ca-77f5-48a8-d1ed-dbe849340eba",
    "papermill": {
     "duration": 0.049495,
     "end_time": "2023-10-28T03:01:49.749213",
     "exception": false,
     "start_time": "2023-10-28T03:01:49.699718",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tensor([[19, 22],\n",
       "        [43, 50]])"
      ]
     },
     "execution_count": 47,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Perform matrix multiplication using the @ operator\n",
    "result_operator = matrix1 @ matrix2\n",
    "result_operator"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "19571f17",
   "metadata": {
    "papermill": {
     "duration": 0.042158,
     "end_time": "2023-10-28T03:01:49.834198",
     "exception": false,
     "start_time": "2023-10-28T03:01:49.792040",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "*PyTorch also provides the @ operator, which is a convenient way to perform matrix multiplication.*"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6327cb30",
   "metadata": {
    "id": "npE1J0hCgJ7o",
    "papermill": {
     "duration": 0.042831,
     "end_time": "2023-10-28T03:01:49.919964",
     "exception": false,
     "start_time": "2023-10-28T03:01:49.877133",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "#### **Important**\n",
    "***One of the most common errors in deep learning (shape errors)\n",
    "Because much of deep learning is multiplying and performing operations on matrices and matrices have a strict rule about what shapes and sizes can be combined, one of the most common errors you'll run into in deep learning is shape mismatches.***"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "id": "76d76294",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:50.007030Z",
     "iopub.status.busy": "2023-10-28T03:01:50.006731Z",
     "iopub.status.idle": "2023-10-28T03:01:50.011912Z",
     "shell.execute_reply": "2023-10-28T03:01:50.011031Z"
    },
    "papermill": {
     "duration": 0.051473,
     "end_time": "2023-10-28T03:01:50.013926",
     "exception": false,
     "start_time": "2023-10-28T03:01:49.962453",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "# Shapes need to be in the right way\n",
    "tensor_A = torch.tensor([[1, 2],\n",
    "                         [3, 4],\n",
    "                         [5, 6]], dtype=torch.float32)\n",
    "\n",
    "tensor_B = torch.tensor([[7, 10],\n",
    "                         [8, 11],\n",
    "                         [9, 12]], dtype=torch.float32)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1d5d1466",
   "metadata": {
    "papermill": {
     "duration": 0.041727,
     "end_time": "2023-10-28T03:01:50.099928",
     "exception": false,
     "start_time": "2023-10-28T03:01:50.058201",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "> ***torch.matmul(tensor_A, tensor_B)***\n",
    "\n",
    "**This will produce an error as expected because of the shape missmatch. the error is as following:**"
   ]
  },
  {
   "attachments": {
    "ff144b9b-048e-478c-af5f-2255abf17f68.png": {
     "image/png": "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"
    }
   },
   "cell_type": "markdown",
   "id": "50f82b9b",
   "metadata": {
    "id": "RDdHuKH2guap",
    "papermill": {
     "duration": 0.040455,
     "end_time": "2023-10-28T03:01:50.180619",
     "exception": false,
     "start_time": "2023-10-28T03:01:50.140164",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "![image.png](attachment:ff144b9b-048e-478c-af5f-2255abf17f68.png)\n",
    "\n",
    "We can make matrix multiplication work between tensor_A and tensor_B by making their inner dimensions match.\n",
    "\n",
    "One of the ways to do this is with a transpose (switch the dimensions of a given tensor).\n",
    "\n",
    "You can perform transposes in PyTorch using either:\n",
    "\n",
    "torch.transpose(input, dim0, dim1) - where input is the desired tensor to transpose and dim0 and dim1 are the dimensions to be swapped.\n",
    "tensor.T - where tensor is the desired tensor to transpose"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "id": "ec0dc1d4",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:50.264951Z",
     "iopub.status.busy": "2023-10-28T03:01:50.264619Z",
     "iopub.status.idle": "2023-10-28T03:01:50.271365Z",
     "shell.execute_reply": "2023-10-28T03:01:50.270241Z"
    },
    "id": "_RNxk2-KgP8w",
    "outputId": "609d78de-067b-4e59-d759-4d2baae77539",
    "papermill": {
     "duration": 0.051171,
     "end_time": "2023-10-28T03:01:50.273539",
     "exception": false,
     "start_time": "2023-10-28T03:01:50.222368",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "tensor([[1., 2.],\n",
      "        [3., 4.],\n",
      "        [5., 6.]])\n",
      "tensor([[ 7., 10.],\n",
      "        [ 8., 11.],\n",
      "        [ 9., 12.]])\n"
     ]
    }
   ],
   "source": [
    "# View tensor_A and tensor_B\n",
    "print(tensor_A)\n",
    "print(tensor_B)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "id": "8433a9ed",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:50.357742Z",
     "iopub.status.busy": "2023-10-28T03:01:50.357457Z",
     "iopub.status.idle": "2023-10-28T03:01:50.364904Z",
     "shell.execute_reply": "2023-10-28T03:01:50.363901Z"
    },
    "id": "k26-z3q4gxl9",
    "outputId": "cbbdeace-e592-4d63-a48c-a1cd5dcde30b",
    "papermill": {
     "duration": 0.051943,
     "end_time": "2023-10-28T03:01:50.366988",
     "exception": false,
     "start_time": "2023-10-28T03:01:50.315045",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "tensor([[1., 2.],\n",
      "        [3., 4.],\n",
      "        [5., 6.]])\n",
      "tensor([[ 7.,  8.,  9.],\n",
      "        [10., 11., 12.]])\n"
     ]
    }
   ],
   "source": [
    "# View tensor_A and tensor_B.T\n",
    "print(tensor_A)\n",
    "print(tensor_B.T)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "id": "dbca2797",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:50.451881Z",
     "iopub.status.busy": "2023-10-28T03:01:50.451604Z",
     "iopub.status.idle": "2023-10-28T03:01:50.469867Z",
     "shell.execute_reply": "2023-10-28T03:01:50.468812Z"
    },
    "id": "LmkQ98JOg0cd",
    "outputId": "ed80521a-1134-4552-e218-6bce8758da78",
    "papermill": {
     "duration": 0.062358,
     "end_time": "2023-10-28T03:01:50.472068",
     "exception": false,
     "start_time": "2023-10-28T03:01:50.409710",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Original shapes: tensor_A = torch.Size([3, 2]), tensor_B = torch.Size([3, 2])\n",
      "\n",
      "New shapes: tensor_A = torch.Size([3, 2]) (same as above), tensor_B.T = torch.Size([2, 3])\n",
      "\n",
      "Multiplying: torch.Size([3, 2]) * torch.Size([2, 3]) <- inner dimensions match\n",
      "\n",
      "Output:\n",
      "\n",
      "tensor([[ 27.,  30.,  33.],\n",
      "        [ 61.,  68.,  75.],\n",
      "        [ 95., 106., 117.]])\n",
      "\n",
      "Output shape: torch.Size([3, 3])\n"
     ]
    }
   ],
   "source": [
    "# The operation works when tensor_B is transposed\n",
    "print(f\"Original shapes: tensor_A = {tensor_A.shape}, tensor_B = {tensor_B.shape}\\n\")\n",
    "print(f\"New shapes: tensor_A = {tensor_A.shape} (same as above), tensor_B.T = {tensor_B.T.shape}\\n\")\n",
    "print(f\"Multiplying: {tensor_A.shape} * {tensor_B.T.shape} <- inner dimensions match\\n\")\n",
    "print(\"Output:\\n\")\n",
    "output = torch.matmul(tensor_A, tensor_B.T)\n",
    "print(output)\n",
    "print(f\"\\nOutput shape: {output.shape}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5d7f62eb",
   "metadata": {
    "id": "0nXo1msahMaa",
    "papermill": {
     "duration": 0.041665,
     "end_time": "2023-10-28T03:01:50.559667",
     "exception": false,
     "start_time": "2023-10-28T03:01:50.518002",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "**You can also use torch.mm() which is a short for torch.matmul().**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "id": "dd77ec3a",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:50.648444Z",
     "iopub.status.busy": "2023-10-28T03:01:50.647998Z",
     "iopub.status.idle": "2023-10-28T03:01:50.656456Z",
     "shell.execute_reply": "2023-10-28T03:01:50.655564Z"
    },
    "id": "r9R5OErEhE8h",
    "outputId": "f88036e2-c861-4b54-994a-c33bafcaa193",
    "papermill": {
     "duration": 0.055695,
     "end_time": "2023-10-28T03:01:50.658371",
     "exception": false,
     "start_time": "2023-10-28T03:01:50.602676",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tensor([[ 27.,  30.,  33.],\n",
       "        [ 61.,  68.,  75.],\n",
       "        [ 95., 106., 117.]])"
      ]
     },
     "execution_count": 52,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# torch.mm is a shortcut for matmul\n",
    "torch.mm(tensor_A, tensor_B.T)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "id": "aebc8634",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:50.744455Z",
     "iopub.status.busy": "2023-10-28T03:01:50.743826Z",
     "iopub.status.idle": "2023-10-28T03:01:50.758999Z",
     "shell.execute_reply": "2023-10-28T03:01:50.757926Z"
    },
    "id": "YECqRh8chPTp",
    "outputId": "834f5224-c47b-4391-fc89-0e5c681b5b8e",
    "papermill": {
     "duration": 0.059028,
     "end_time": "2023-10-28T03:01:50.761021",
     "exception": false,
     "start_time": "2023-10-28T03:01:50.701993",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Input shape: torch.Size([3, 2])\n",
      "\n",
      "Output:\n",
      "tensor([[2.2368, 1.2292, 0.4714, 0.3864, 0.1309, 0.9838],\n",
      "        [4.4919, 2.1970, 0.4469, 0.5285, 0.3401, 2.4777],\n",
      "        [6.7469, 3.1648, 0.4224, 0.6705, 0.5493, 3.9716]],\n",
      "       grad_fn=<AddmmBackward0>)\n",
      "\n",
      "Output shape: torch.Size([3, 6])\n"
     ]
    }
   ],
   "source": [
    "# Since the linear layer starts with a random weights matrix, let's make it reproducible (more on this later)\n",
    "torch.manual_seed(42)\n",
    "# This uses matrix multiplication\n",
    "linear = torch.nn.Linear(in_features=2, # in_features = matches inner dimension of input\n",
    "                         out_features=6) # out_features = describes outer value\n",
    "x = tensor_A\n",
    "output = linear(x)\n",
    "print(f\"Input shape: {x.shape}\\n\")\n",
    "print(f\"Output:\\n{output}\\n\\nOutput shape: {output.shape}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "id": "61277087",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:50.848032Z",
     "iopub.status.busy": "2023-10-28T03:01:50.847770Z",
     "iopub.status.idle": "2023-10-28T03:01:50.855059Z",
     "shell.execute_reply": "2023-10-28T03:01:50.854297Z"
    },
    "id": "fzCA3oe-jHmx",
    "outputId": "b2733b11-d9e4-4199-ca39-c22250396f8b",
    "papermill": {
     "duration": 0.051898,
     "end_time": "2023-10-28T03:01:50.856901",
     "exception": false,
     "start_time": "2023-10-28T03:01:50.805003",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(tensor(6.), tensor(1.), tensor(3.5000), tensor(21.))"
      ]
     },
     "execution_count": 54,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "torch.max(x), torch.min(x), torch.mean(x.type(torch.float32)), torch.sum(x)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "bb9c9f9c",
   "metadata": {
    "id": "eE7QxSbTjZbm",
    "papermill": {
     "duration": 0.041958,
     "end_time": "2023-10-28T03:01:50.941780",
     "exception": false,
     "start_time": "2023-10-28T03:01:50.899822",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "## **Positional Min/Max**\n",
    "You can use torch.argmax() and torch.argmin() functions to find the indices where the maximum and minimum values occur in a tensor, respectively. This is useful when you need to determine the positions of extreme values without necessarily retrieving the actual values. We will explore their practical applications further, especially when working with the softmax activation function in a later section.\n",
    "\n",
    "Here's a concise explanation of how to use these functions:\n",
    "\n",
    "- **torch.argmax(tensor)** returns the index of the maximum value in the tensor.\n",
    "- **torch.argmin(tensor)** returns the index of the minimum value in the tensor."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "id": "401a62ee",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:51.070666Z",
     "iopub.status.busy": "2023-10-28T03:01:51.070211Z",
     "iopub.status.idle": "2023-10-28T03:01:51.080420Z",
     "shell.execute_reply": "2023-10-28T03:01:51.079417Z"
    },
    "id": "iyVGZm-fjn2U",
    "outputId": "20751042-8b4f-4d60-fcdd-d60f0e3a7c3f",
    "papermill": {
     "duration": 0.099364,
     "end_time": "2023-10-28T03:01:51.082247",
     "exception": false,
     "start_time": "2023-10-28T03:01:50.982883",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Tensor: tensor([10, 20, 30, 40, 50, 60, 70, 80, 90])\n",
      "Index where max value occurs: 8\n",
      "Index where min value occurs: 0\n"
     ]
    }
   ],
   "source": [
    "# Create a tensor\n",
    "tensor = torch.arange(10, 100, 10)\n",
    "print(f\"Tensor: {tensor}\")\n",
    "\n",
    "# Returns index of max and min values\n",
    "print(f\"Index where max value occurs: {tensor.argmax()}\")\n",
    "print(f\"Index where min value occurs: {tensor.argmin()}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d32f662e",
   "metadata": {
    "id": "oUcak_mwj2YC",
    "papermill": {
     "duration": 0.045124,
     "end_time": "2023-10-28T03:01:51.170591",
     "exception": false,
     "start_time": "2023-10-28T03:01:51.125467",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "## **Change Tensor Datatype**\n",
    "\n",
    "If we have one tensor in torch.float64 and another in torch.float32, we might encounter data type compatibility issues when performing certain operations. However, we can address this by adjusting the data type of the tensors using torch.Tensor.type(dtype=None), where the dtype parameter specifies the desired data type.\n",
    "\n",
    "Let's start by creating a tensor and examining its data type (the default is torch.float32)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "id": "e9ef8c98",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:51.256691Z",
     "iopub.status.busy": "2023-10-28T03:01:51.256355Z",
     "iopub.status.idle": "2023-10-28T03:01:51.263852Z",
     "shell.execute_reply": "2023-10-28T03:01:51.262962Z"
    },
    "id": "8DcVoW8nkBH9",
    "outputId": "9d71d456-3e1b-4b17-fbfb-3a5d918f23e8",
    "papermill": {
     "duration": 0.053493,
     "end_time": "2023-10-28T03:01:51.265939",
     "exception": false,
     "start_time": "2023-10-28T03:01:51.212446",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(torch.float32, tensor([10., 20., 30., 40., 50., 60., 70., 80., 90.]))"
      ]
     },
     "execution_count": 56,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Create a tensor and check its datatype\n",
    "sample_tensor = torch.arange(10., 100., 10.)\n",
    "sample_tensor.dtype, sample_tensor"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "id": "e25da044",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:51.370210Z",
     "iopub.status.busy": "2023-10-28T03:01:51.369841Z",
     "iopub.status.idle": "2023-10-28T03:01:51.376984Z",
     "shell.execute_reply": "2023-10-28T03:01:51.376083Z"
    },
    "id": "49reLBh8kDmk",
    "outputId": "355ef9b2-8d3e-451d-8fcb-c5ced2bf09d8",
    "papermill": {
     "duration": 0.065196,
     "end_time": "2023-10-28T03:01:51.378999",
     "exception": false,
     "start_time": "2023-10-28T03:01:51.313803",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tensor([10., 20., 30., 40., 50., 60., 70., 80., 90.], dtype=torch.float16)"
      ]
     },
     "execution_count": 57,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Create a float16 tensor\n",
    "tensor_float16 = sample_tensor.type(torch.float16)\n",
    "tensor_float16"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "id": "bc748a4a",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:51.464408Z",
     "iopub.status.busy": "2023-10-28T03:01:51.464081Z",
     "iopub.status.idle": "2023-10-28T03:01:51.470690Z",
     "shell.execute_reply": "2023-10-28T03:01:51.469801Z"
    },
    "id": "fxepDnVGkGbk",
    "outputId": "87d95193-62e4-48ba-d212-de4b5a89b548",
    "papermill": {
     "duration": 0.051667,
     "end_time": "2023-10-28T03:01:51.472649",
     "exception": false,
     "start_time": "2023-10-28T03:01:51.420982",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tensor([10, 20, 30, 40, 50, 60, 70, 80, 90], dtype=torch.int8)"
      ]
     },
     "execution_count": 58,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Create a int8 tensor\n",
    "tensor_int8 = sample_tensor.type(torch.int8)\n",
    "tensor_int8"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2bbcfd23",
   "metadata": {
    "id": "2P3zULJGkZVs",
    "papermill": {
     "duration": 0.043297,
     "end_time": "2023-10-28T03:01:51.559159",
     "exception": false,
     "start_time": "2023-10-28T03:01:51.515862",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "## **Reshaping, Stacking, Squeezing and Unsqueezing**\n",
    "Often times you'll want to reshape or change the dimensions of your tensors without actually changing the values ins\"ide them.\n",
    "\n",
    "To do so, some popular methods are:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "id": "aade50d9",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:51.648654Z",
     "iopub.status.busy": "2023-10-28T03:01:51.648299Z",
     "iopub.status.idle": "2023-10-28T03:01:51.655689Z",
     "shell.execute_reply": "2023-10-28T03:01:51.654700Z"
    },
    "id": "FCUUTTnak8_O",
    "outputId": "a3c818b7-e35b-418e-9027-151b41e6a96a",
    "papermill": {
     "duration": 0.054198,
     "end_time": "2023-10-28T03:01:51.657882",
     "exception": false,
     "start_time": "2023-10-28T03:01:51.603684",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(tensor([1., 2., 3., 4., 5., 6., 7., 8., 9.]), torch.Size([9]))"
      ]
     },
     "execution_count": 59,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Create a Tensor\n",
    "x = torch.arange(1., 10.)\n",
    "x, x.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "id": "c22e1b85",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:51.744679Z",
     "iopub.status.busy": "2023-10-28T03:01:51.744410Z",
     "iopub.status.idle": "2023-10-28T03:01:51.750272Z",
     "shell.execute_reply": "2023-10-28T03:01:51.749314Z"
    },
    "id": "FPiV3ieSsZIw",
    "outputId": "4bdca098-b6c1-4319-ddcc-35e172231daa",
    "papermill": {
     "duration": 0.051666,
     "end_time": "2023-10-28T03:01:51.752339",
     "exception": false,
     "start_time": "2023-10-28T03:01:51.700673",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Reshaped Tensor:\n",
      "tensor([[1., 2., 3., 4., 5., 6., 7., 8., 9.]])\n",
      "Shape of Reshaped Tensor: torch.Size([1, 9])\n"
     ]
    }
   ],
   "source": [
    "# Add an extra dimention\n",
    "x_reshaped = x.reshape(1, 9)\n",
    "\n",
    "print(\"Reshaped Tensor:\")\n",
    "print(x_reshaped)\n",
    "print(\"Shape of Reshaped Tensor:\", x_reshaped.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "id": "445ffd6a",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:51.838159Z",
     "iopub.status.busy": "2023-10-28T03:01:51.837894Z",
     "iopub.status.idle": "2023-10-28T03:01:51.843991Z",
     "shell.execute_reply": "2023-10-28T03:01:51.843026Z"
    },
    "id": "RGn8b_4vtIJQ",
    "outputId": "ca4dc11c-6638-4ea4-d081-254830cb917e",
    "papermill": {
     "duration": 0.051548,
     "end_time": "2023-10-28T03:01:51.845924",
     "exception": false,
     "start_time": "2023-10-28T03:01:51.794376",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Reshaped Tensor:\n",
      "tensor([[1.],\n",
      "        [2.],\n",
      "        [3.],\n",
      "        [4.],\n",
      "        [5.],\n",
      "        [6.],\n",
      "        [7.],\n",
      "        [8.],\n",
      "        [9.]])\n",
      "Shape of Reshaped Tensor: torch.Size([9, 1])\n"
     ]
    }
   ],
   "source": [
    "# Add an extra dimension using reshape\n",
    "x_reshaped = x.reshape(9, 1)\n",
    "\n",
    "print(\"Reshaped Tensor:\")\n",
    "print(x_reshaped)\n",
    "print(\"Shape of Reshaped Tensor:\", x_reshaped.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "id": "51149aef",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:51.932786Z",
     "iopub.status.busy": "2023-10-28T03:01:51.932510Z",
     "iopub.status.idle": "2023-10-28T03:01:51.938482Z",
     "shell.execute_reply": "2023-10-28T03:01:51.937531Z"
    },
    "id": "UnoMZk5ksimQ",
    "outputId": "c9916768-6143-45cc-cc59-788c6dc22a5a",
    "papermill": {
     "duration": 0.05179,
     "end_time": "2023-10-28T03:01:51.940449",
     "exception": false,
     "start_time": "2023-10-28T03:01:51.888659",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Modified View Tensor:\n",
      "tensor([[1., 2., 3., 4., 5., 6., 7., 8., 9.]])\n",
      "Shape of Modified View Tensor: torch.Size([1, 9])\n"
     ]
    }
   ],
   "source": [
    "# Change the view using .view()\n",
    "z = x.view(1, 9)\n",
    "\n",
    "print(\"Modified View Tensor:\")\n",
    "print(z)\n",
    "print(\"Shape of Modified View Tensor:\", z.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4c29e44b",
   "metadata": {
    "id": "I_hF0L5lujMW",
    "papermill": {
     "duration": 0.041826,
     "end_time": "2023-10-28T03:01:52.025399",
     "exception": false,
     "start_time": "2023-10-28T03:01:51.983573",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "*Remember though, changing the view of a tensor with torch.view() really only creates a new view of the same tensor.So changing the view changes the original tensor too.*"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "id": "7d969afe",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:52.115767Z",
     "iopub.status.busy": "2023-10-28T03:01:52.115426Z",
     "iopub.status.idle": "2023-10-28T03:01:52.122529Z",
     "shell.execute_reply": "2023-10-28T03:01:52.121505Z"
    },
    "id": "dzHfHtS9vFh3",
    "outputId": "2266034d-8608-4411-eeee-1e4bea66ff6e",
    "papermill": {
     "duration": 0.055085,
     "end_time": "2023-10-28T03:01:52.124457",
     "exception": false,
     "start_time": "2023-10-28T03:01:52.069372",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Modified View Tensor z:\n",
      "tensor([[5., 2., 3., 4., 5., 6., 7., 8., 9.]])\n",
      "\n",
      "Original Tensor x:\n",
      "tensor([5., 2., 3., 4., 5., 6., 7., 8., 9.])\n"
     ]
    }
   ],
   "source": [
    "# Changing z changes x\n",
    "z[:, 0] = 5\n",
    "\n",
    "print(\"Modified View Tensor z:\")\n",
    "print(z)\n",
    "print(\"\\nOriginal Tensor x:\")\n",
    "print(x)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "703d95d7",
   "metadata": {
    "id": "s_MLgW7VvJl8",
    "papermill": {
     "duration": 0.042366,
     "end_time": "2023-10-28T03:01:52.210388",
     "exception": false,
     "start_time": "2023-10-28T03:01:52.168022",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "\n",
    "If we wanted to stack our new tensor on top of itself five times, we could do so with torch.stack()."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "id": "ee43a395",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:52.298049Z",
     "iopub.status.busy": "2023-10-28T03:01:52.297743Z",
     "iopub.status.idle": "2023-10-28T03:01:52.306290Z",
     "shell.execute_reply": "2023-10-28T03:01:52.305430Z"
    },
    "id": "bEeohO8DvPbD",
    "outputId": "b054482a-96ee-4e01-c795-3aa375c6b3fe",
    "papermill": {
     "duration": 0.054714,
     "end_time": "2023-10-28T03:01:52.308196",
     "exception": false,
     "start_time": "2023-10-28T03:01:52.253482",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tensor([[1, 2, 3, 4, 5, 6, 7, 8, 9],\n",
       "        [1, 2, 3, 4, 5, 6, 7, 8, 9],\n",
       "        [1, 2, 3, 4, 5, 6, 7, 8, 9],\n",
       "        [1, 2, 3, 4, 5, 6, 7, 8, 9]])"
      ]
     },
     "execution_count": 64,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Original tensor\n",
    "x = torch.tensor([1, 2, 3, 4, 5, 6, 7, 8, 9])\n",
    "\n",
    "# Stack tensors on top of each other along rows (dim=0)\n",
    "x_stacked = torch.stack([x, x, x, x], dim=0)\n",
    "\n",
    "# Stacked tensor\n",
    "x_stacked"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "id": "8c64b1c4",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:52.393245Z",
     "iopub.status.busy": "2023-10-28T03:01:52.392985Z",
     "iopub.status.idle": "2023-10-28T03:01:52.399915Z",
     "shell.execute_reply": "2023-10-28T03:01:52.399062Z"
    },
    "id": "6x7UPXzuvc0c",
    "outputId": "f12e8d44-2b82-4f14-edba-eadf76414bfa",
    "papermill": {
     "duration": 0.051368,
     "end_time": "2023-10-28T03:01:52.401734",
     "exception": false,
     "start_time": "2023-10-28T03:01:52.350366",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tensor([[1, 1, 1, 1],\n",
       "        [2, 2, 2, 2],\n",
       "        [3, 3, 3, 3],\n",
       "        [4, 4, 4, 4],\n",
       "        [5, 5, 5, 5],\n",
       "        [6, 6, 6, 6],\n",
       "        [7, 7, 7, 7],\n",
       "        [8, 8, 8, 8],\n",
       "        [9, 9, 9, 9]])"
      ]
     },
     "execution_count": 65,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Stack tensors on top of each other along columns (dim=1)\n",
    "x_stacked = torch.stack([x, x, x, x], dim=1)\n",
    "\n",
    "# Display the stacked tensor\n",
    "x_stacked"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c4f8be5e",
   "metadata": {
    "id": "7LiqpMolvkzv",
    "papermill": {
     "duration": 0.04214,
     "end_time": "2023-10-28T03:01:52.485998",
     "exception": false,
     "start_time": "2023-10-28T03:01:52.443858",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "## **Removing Single Dimention**\n",
    "\n",
    "#### How about removing all single dimensions from a tensor?\n",
    "\n",
    "To do so you can use torch.squeeze() (I remember this as squeezing the tensor to only have dimensions over 1)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "id": "893cd633",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:52.573447Z",
     "iopub.status.busy": "2023-10-28T03:01:52.572457Z",
     "iopub.status.idle": "2023-10-28T03:01:52.578989Z",
     "shell.execute_reply": "2023-10-28T03:01:52.578151Z"
    },
    "id": "2fW7lOq1v9Nk",
    "outputId": "05ca7209-0e24-4cde-eb65-624a51aa70ce",
    "papermill": {
     "duration": 0.052133,
     "end_time": "2023-10-28T03:01:52.580857",
     "exception": false,
     "start_time": "2023-10-28T03:01:52.528724",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tensor([[5.],\n",
       "        [2.],\n",
       "        [3.],\n",
       "        [4.],\n",
       "        [5.],\n",
       "        [6.],\n",
       "        [7.],\n",
       "        [8.],\n",
       "        [9.]])"
      ]
     },
     "execution_count": 66,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x_reshaped"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "id": "001e3ca3",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:52.668826Z",
     "iopub.status.busy": "2023-10-28T03:01:52.668544Z",
     "iopub.status.idle": "2023-10-28T03:01:52.674970Z",
     "shell.execute_reply": "2023-10-28T03:01:52.674109Z"
    },
    "id": "uWTkqB6yv9Fd",
    "outputId": "865497c6-b176-4a16-8257-38ccae03716c",
    "papermill": {
     "duration": 0.053053,
     "end_time": "2023-10-28T03:01:52.676810",
     "exception": false,
     "start_time": "2023-10-28T03:01:52.623757",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tensor([[5., 2., 3., 4., 5., 6., 7., 8., 9.]])"
      ]
     },
     "execution_count": 67,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Reshaping x_reshaped \n",
    "x_reshaped = x_reshaped.reshape(1, 9)\n",
    "x_reshaped"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d656f26a",
   "metadata": {
    "papermill": {
     "duration": 0.043671,
     "end_time": "2023-10-28T03:01:52.762730",
     "exception": false,
     "start_time": "2023-10-28T03:01:52.719059",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "***When you use the squeeze() method on a tensor, it removes dimensions with size 1, effectively reducing the rank (number of dimensions) of the tensor. Let's apply squeeze() to the x_reshaped tensor:***"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "id": "c9a03a2d",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:52.852280Z",
     "iopub.status.busy": "2023-10-28T03:01:52.851943Z",
     "iopub.status.idle": "2023-10-28T03:01:52.858499Z",
     "shell.execute_reply": "2023-10-28T03:01:52.857731Z"
    },
    "id": "mHorWYB3wZiu",
    "outputId": "8d6b6216-fc54-40af-ea96-3fe2a250ba0d",
    "papermill": {
     "duration": 0.053302,
     "end_time": "2023-10-28T03:01:52.860404",
     "exception": false,
     "start_time": "2023-10-28T03:01:52.807102",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tensor([5., 2., 3., 4., 5., 6., 7., 8., 9.])"
      ]
     },
     "execution_count": 68,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Squeeze the x_reshaped tensor\n",
    "x_reshaped_squeezed = x_reshaped.squeeze()\n",
    "\n",
    "# Display the squeezed tensor\n",
    "x_reshaped_squeezed"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "id": "0aa8f43d",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:52.949501Z",
     "iopub.status.busy": "2023-10-28T03:01:52.949202Z",
     "iopub.status.idle": "2023-10-28T03:01:52.954934Z",
     "shell.execute_reply": "2023-10-28T03:01:52.954066Z"
    },
    "id": "XlxWbRwpwh6l",
    "outputId": "e899528a-b7b1-4093-ef68-498fe6ea4245",
    "papermill": {
     "duration": 0.052818,
     "end_time": "2023-10-28T03:01:52.956799",
     "exception": false,
     "start_time": "2023-10-28T03:01:52.903981",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "torch.Size([9])"
      ]
     },
     "execution_count": 69,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x_reshaped.squeeze().size()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "id": "11328515",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:53.048155Z",
     "iopub.status.busy": "2023-10-28T03:01:53.047850Z",
     "iopub.status.idle": "2023-10-28T03:01:53.055147Z",
     "shell.execute_reply": "2023-10-28T03:01:53.054008Z"
    },
    "id": "GXVReIqIwlUk",
    "outputId": "2ebda211-d109-4be3-fe75-fc246769f9ef",
    "papermill": {
     "duration": 0.055825,
     "end_time": "2023-10-28T03:01:53.057148",
     "exception": false,
     "start_time": "2023-10-28T03:01:53.001323",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Previous tensor: tensor([[5., 2., 3., 4., 5., 6., 7., 8., 9.]])\n",
      "Previous shape: torch.Size([1, 9])\n",
      "\n",
      "New tensor: tensor([5., 2., 3., 4., 5., 6., 7., 8., 9.])\n",
      "New shape: torch.Size([9])\n"
     ]
    }
   ],
   "source": [
    "print(f\"Previous tensor: {x_reshaped}\")\n",
    "print(f\"Previous shape: {x_reshaped.shape}\")\n",
    "\n",
    "# Remove extra dimension from x_reshaped\n",
    "x_squeezed = x_reshaped.squeeze()\n",
    "\n",
    "print(f\"\\nNew tensor: {x_squeezed}\")\n",
    "print(f\"New shape: {x_squeezed.shape}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "08e4fd33",
   "metadata": {
    "id": "f097hnGpwrbE",
    "papermill": {
     "duration": 0.04318,
     "end_time": "2023-10-28T03:01:53.143940",
     "exception": false,
     "start_time": "2023-10-28T03:01:53.100760",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "### **And to do the reverse of torch.squeeze() you can use torch.unsqueeze() to add a dimension value of 1 at a specific index.**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "id": "74705f93",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:53.235678Z",
     "iopub.status.busy": "2023-10-28T03:01:53.234973Z",
     "iopub.status.idle": "2023-10-28T03:01:53.243120Z",
     "shell.execute_reply": "2023-10-28T03:01:53.242224Z"
    },
    "id": "mUEbfAMlxCOn",
    "outputId": "eb8cdd3c-4726-4bf5-b09b-df8c57c56d86",
    "papermill": {
     "duration": 0.057363,
     "end_time": "2023-10-28T03:01:53.245073",
     "exception": false,
     "start_time": "2023-10-28T03:01:53.187710",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Previous tensor: tensor([5., 2., 3., 4., 5., 6., 7., 8., 9.])\n",
      "Previous shape: torch.Size([9])\n",
      "\n",
      "New tensor: tensor([[5., 2., 3., 4., 5., 6., 7., 8., 9.]])\n",
      "New shape: torch.Size([1, 9])\n"
     ]
    }
   ],
   "source": [
    "print(f\"Previous tensor: {x_squeezed}\")\n",
    "print(f\"Previous shape: {x_squeezed.shape}\")\n",
    "\n",
    "## Add an extra dimension with unsqueeze\n",
    "x_unsqueezed = x_squeezed.unsqueeze(dim=0)\n",
    "print(f\"\\nNew tensor: {x_unsqueezed}\")\n",
    "print(f\"New shape: {x_unsqueezed.shape}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "44706b58",
   "metadata": {
    "id": "M6xZ-JX_xGa7",
    "papermill": {
     "duration": 0.04533,
     "end_time": "2023-10-28T03:01:53.336960",
     "exception": false,
     "start_time": "2023-10-28T03:01:53.291630",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "### **Another essential tensor operation is rearranging the order of axes values, which can be achieved using torch.permute(input, dims). This function allows you to create a view of the input tensor with the specified dimensions, effectively rearranging the axes.**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "id": "6a9c9442",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:53.428202Z",
     "iopub.status.busy": "2023-10-28T03:01:53.427377Z",
     "iopub.status.idle": "2023-10-28T03:01:53.435340Z",
     "shell.execute_reply": "2023-10-28T03:01:53.434298Z"
    },
    "id": "nELYm077xkY0",
    "outputId": "56dcef55-9dd8-450b-e155-aa655a293f1a",
    "papermill": {
     "duration": 0.055627,
     "end_time": "2023-10-28T03:01:53.437629",
     "exception": false,
     "start_time": "2023-10-28T03:01:53.382002",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Previous shape: torch.Size([224, 224, 3])\n",
      "New shape: torch.Size([3, 224, 224])\n"
     ]
    }
   ],
   "source": [
    "# Create tensor with specific shape\n",
    "x_original = torch.rand(size=(224, 224, 3)) # Height, Width, Channels\n",
    "\n",
    "# Permute the original tensor to rearrange the axis order\n",
    "x_permuted = x_original.permute(2, 0, 1) # shifts axis 0->1, 1->2, 2->0\n",
    "\n",
    "print(f\"Previous shape: {x_original.shape}\")\n",
    "print(f\"New shape: {x_permuted.shape}\") # Channels, Height, Width"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "id": "ca085caf",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:53.528778Z",
     "iopub.status.busy": "2023-10-28T03:01:53.528469Z",
     "iopub.status.idle": "2023-10-28T03:01:53.535142Z",
     "shell.execute_reply": "2023-10-28T03:01:53.534275Z"
    },
    "id": "zptkeqbxxzw8",
    "outputId": "27a874c5-afeb-494c-e05d-930850376b24",
    "papermill": {
     "duration": 0.055361,
     "end_time": "2023-10-28T03:01:53.537117",
     "exception": false,
     "start_time": "2023-10-28T03:01:53.481756",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "torch.Size([2, 3, 5])"
      ]
     },
     "execution_count": 73,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Create a 3D tensor with random values\n",
    "x = torch.randn(2, 3, 5)\n",
    "\n",
    "# Use the .size() method to get the size (shape) of the tensor\n",
    "size_of_x = x.size()\n",
    "size_of_x"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "id": "78ca460a",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:53.626174Z",
     "iopub.status.busy": "2023-10-28T03:01:53.625882Z",
     "iopub.status.idle": "2023-10-28T03:01:53.632129Z",
     "shell.execute_reply": "2023-10-28T03:01:53.631447Z"
    },
    "id": "yIRlTuB8yEM8",
    "outputId": "01c498d9-b844-42eb-a1c3-aa6a74cba686",
    "papermill": {
     "duration": 0.052895,
     "end_time": "2023-10-28T03:01:53.634026",
     "exception": false,
     "start_time": "2023-10-28T03:01:53.581131",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tensor([[[-0.7963,  0.6134,  1.1794, -0.9621,  1.6388],\n",
       "         [-0.8928,  1.2756, -0.9238,  1.6126, -0.7292],\n",
       "         [ 0.7680,  2.4192, -0.4922, -0.6502,  0.4609]],\n",
       "\n",
       "        [[-0.5408, -0.3830, -1.1725,  0.0145,  0.5402],\n",
       "         [ 1.4513,  0.4064,  0.9798, -2.6272, -1.1485],\n",
       "         [-0.3889,  1.5624, -0.7264,  1.0032,  0.4538]]])"
      ]
     },
     "execution_count": 74,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "id": "397550ef",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:53.721910Z",
     "iopub.status.busy": "2023-10-28T03:01:53.721601Z",
     "iopub.status.idle": "2023-10-28T03:01:53.728183Z",
     "shell.execute_reply": "2023-10-28T03:01:53.727329Z"
    },
    "id": "gVMAtlNSyIBc",
    "outputId": "bec5f42b-7ea9-4c7d-9f67-d29bcb61c953",
    "papermill": {
     "duration": 0.052831,
     "end_time": "2023-10-28T03:01:53.730171",
     "exception": false,
     "start_time": "2023-10-28T03:01:53.677340",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "torch.Size([5, 2, 3])"
      ]
     },
     "execution_count": 75,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Permute the dimensions of tensor x using torch.permute()\n",
    "# The specified permutation order is (2, 0, 1)\n",
    "permuted_x = torch.permute(x, (2, 0, 1))\n",
    "\n",
    "# Get the size (shape) of the permuted tensor\n",
    "size_of_permuted_x = permuted_x.size()\n",
    "\n",
    "# Display the size of the permuted tensor\n",
    "size_of_permuted_x"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "id": "edaae8ae",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:53.825847Z",
     "iopub.status.busy": "2023-10-28T03:01:53.825075Z",
     "iopub.status.idle": "2023-10-28T03:01:53.831320Z",
     "shell.execute_reply": "2023-10-28T03:01:53.830489Z"
    },
    "id": "QENeivz-yRNT",
    "outputId": "678033e1-cfda-4400-910f-6047710d31e1",
    "papermill": {
     "duration": 0.056776,
     "end_time": "2023-10-28T03:01:53.833366",
     "exception": false,
     "start_time": "2023-10-28T03:01:53.776590",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tensor(0.2666)"
      ]
     },
     "execution_count": 76,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x_original[0,0,0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "id": "80d6b96b",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:53.923475Z",
     "iopub.status.busy": "2023-10-28T03:01:53.922984Z",
     "iopub.status.idle": "2023-10-28T03:01:53.929020Z",
     "shell.execute_reply": "2023-10-28T03:01:53.928265Z"
    },
    "id": "QCnulSTPzBID",
    "outputId": "7ca27bef-8778-4abf-ebaa-319546a633d4",
    "papermill": {
     "duration": 0.054129,
     "end_time": "2023-10-28T03:01:53.930930",
     "exception": false,
     "start_time": "2023-10-28T03:01:53.876801",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tensor(0.2666)"
      ]
     },
     "execution_count": 77,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x_permuted[0,0,0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "id": "59ec635e",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:54.023830Z",
     "iopub.status.busy": "2023-10-28T03:01:54.023543Z",
     "iopub.status.idle": "2023-10-28T03:01:54.030972Z",
     "shell.execute_reply": "2023-10-28T03:01:54.030129Z"
    },
    "id": "SPhIFZ5dzDpr",
    "outputId": "0a348eec-02af-4717-cc68-c9dded9c3d84",
    "papermill": {
     "duration": 0.056086,
     "end_time": "2023-10-28T03:01:54.032928",
     "exception": false,
     "start_time": "2023-10-28T03:01:53.976842",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(tensor(0.5500), tensor(0.5500))"
      ]
     },
     "execution_count": 78,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Modify the value at a specific position in the original tensor x_original\n",
    "x_original[0, 0, 0] = 0.55\n",
    "\n",
    "# Check the values at the same position in both the original and permuted tensors\n",
    "value_in_original = x_original[0, 0, 0]\n",
    "value_in_permuted = x_permuted[0, 0, 0]\n",
    "\n",
    "value_in_original, value_in_permuted"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7800065e",
   "metadata": {
    "papermill": {
     "duration": 0.043843,
     "end_time": "2023-10-28T03:01:54.123081",
     "exception": false,
     "start_time": "2023-10-28T03:01:54.079238",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "*- We first modify the value at position (0, 0, 0) in the original tensor x_original and set it to 0.55.*\n",
    "\n",
    "*- Then, we retrieve and store the values at the same position in both the original and permuted tensors.*\n",
    "\n",
    "*- Finally, we display the values.*"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9d4bb6bf",
   "metadata": {
    "id": "9Gu8O0rdzSwD",
    "papermill": {
     "duration": 0.043274,
     "end_time": "2023-10-28T03:01:54.211246",
     "exception": false,
     "start_time": "2023-10-28T03:01:54.167972",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "## **Indexing**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "id": "6f547dbf",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:54.307392Z",
     "iopub.status.busy": "2023-10-28T03:01:54.307018Z",
     "iopub.status.idle": "2023-10-28T03:01:54.314486Z",
     "shell.execute_reply": "2023-10-28T03:01:54.313479Z"
    },
    "id": "n99mTlJ4znut",
    "outputId": "92f2694d-7798-46ea-a554-363bd1da8fb8",
    "papermill": {
     "duration": 0.059561,
     "end_time": "2023-10-28T03:01:54.316542",
     "exception": false,
     "start_time": "2023-10-28T03:01:54.256981",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(tensor([[[1, 2, 3],\n",
       "          [4, 5, 6],\n",
       "          [7, 8, 9]]]),\n",
       " torch.Size([1, 3, 3]))"
      ]
     },
     "execution_count": 79,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Create a tensor using torch.arange() and reshape it\n",
    "x = torch.arange(1, 10).reshape(1, 3, 3)\n",
    "\n",
    "# Display the tensor x and its shape\n",
    "x, x.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "id": "8d0f6785",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:54.412113Z",
     "iopub.status.busy": "2023-10-28T03:01:54.411577Z",
     "iopub.status.idle": "2023-10-28T03:01:54.417830Z",
     "shell.execute_reply": "2023-10-28T03:01:54.416989Z"
    },
    "id": "PSO3jShkzzAz",
    "outputId": "8e594e67-2f28-457b-9e37-4b9a9949ff5a",
    "papermill": {
     "duration": 0.056181,
     "end_time": "2023-10-28T03:01:54.419839",
     "exception": false,
     "start_time": "2023-10-28T03:01:54.363658",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tensor([[1, 2, 3],\n",
       "        [4, 5, 6],\n",
       "        [7, 8, 9]])"
      ]
     },
     "execution_count": 80,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Access the first element along the first dimension of the tensor x\n",
    "first_element = x[0]\n",
    "\n",
    "# Printing the first element (a 2D slice) of the tensor x\n",
    "first_element"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "id": "5b55210e",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:54.514383Z",
     "iopub.status.busy": "2023-10-28T03:01:54.513378Z",
     "iopub.status.idle": "2023-10-28T03:01:54.520020Z",
     "shell.execute_reply": "2023-10-28T03:01:54.519147Z"
    },
    "id": "NrLKjc99z7Ii",
    "outputId": "49bd7809-460b-4a97-e92f-52ccddbc820e",
    "papermill": {
     "duration": 0.056221,
     "end_time": "2023-10-28T03:01:54.522082",
     "exception": false,
     "start_time": "2023-10-28T03:01:54.465861",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tensor([1, 2, 3])"
      ]
     },
     "execution_count": 81,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Access the element at the first row and first column of the first 2D matrix in the tensor x\n",
    "element = x[0][0]\n",
    "\n",
    "# Printing the accessed element\n",
    "element"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "id": "d6ca4c1e",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:54.613098Z",
     "iopub.status.busy": "2023-10-28T03:01:54.612819Z",
     "iopub.status.idle": "2023-10-28T03:01:54.618788Z",
     "shell.execute_reply": "2023-10-28T03:01:54.617987Z"
    },
    "id": "_jGgtTjZz_4x",
    "outputId": "7ba5113f-cc24-495f-899b-6faa59d14f1f",
    "papermill": {
     "duration": 0.05353,
     "end_time": "2023-10-28T03:01:54.620662",
     "exception": false,
     "start_time": "2023-10-28T03:01:54.567132",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tensor([1, 2, 3])"
      ]
     },
     "execution_count": 82,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Access the element at the first row and first column of the first 2D matrix in the tensor x\n",
    "element = x[0, 0]\n",
    "element"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "id": "591a8bd3",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:54.710586Z",
     "iopub.status.busy": "2023-10-28T03:01:54.710104Z",
     "iopub.status.idle": "2023-10-28T03:01:54.716109Z",
     "shell.execute_reply": "2023-10-28T03:01:54.715302Z"
    },
    "id": "cRnU34AA0DVQ",
    "outputId": "de8f5a57-7aa3-4064-ed01-37d17affc484",
    "papermill": {
     "duration": 0.052723,
     "end_time": "2023-10-28T03:01:54.718071",
     "exception": false,
     "start_time": "2023-10-28T03:01:54.665348",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tensor(1)"
      ]
     },
     "execution_count": 83,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Access the element at the first row, first column, and first depth dimension of the tensor x\n",
    "element = x[0, 0, 0]\n",
    "element"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "id": "8d172cec",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:54.809011Z",
     "iopub.status.busy": "2023-10-28T03:01:54.808767Z",
     "iopub.status.idle": "2023-10-28T03:01:54.814739Z",
     "shell.execute_reply": "2023-10-28T03:01:54.813847Z"
    },
    "id": "x8yoDlbg0M37",
    "outputId": "9e7aec5e-3bea-42b0-fcae-56ed766ef1c4",
    "papermill": {
     "duration": 0.053494,
     "end_time": "2023-10-28T03:01:54.816653",
     "exception": false,
     "start_time": "2023-10-28T03:01:54.763159",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tensor(2)"
      ]
     },
     "execution_count": 84,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Access the element at the first row, first column, and second depth dimension of the tensor x\n",
    "element = x[0, 0, 1]\n",
    "element"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "id": "db8c6726",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:54.908060Z",
     "iopub.status.busy": "2023-10-28T03:01:54.907781Z",
     "iopub.status.idle": "2023-10-28T03:01:54.913775Z",
     "shell.execute_reply": "2023-10-28T03:01:54.913075Z"
    },
    "id": "-DbrLVUB0Ytp",
    "outputId": "65d9d4b9-cfcb-4537-bcce-9e017d9f2277",
    "papermill": {
     "duration": 0.054415,
     "end_time": "2023-10-28T03:01:54.915574",
     "exception": false,
     "start_time": "2023-10-28T03:01:54.861159",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tensor(5)"
      ]
     },
     "execution_count": 85,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Access the element at the first row, second column, and second depth dimension of the tensor x\n",
    "x[0, 1, 1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "id": "f0f84f1c",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:55.005690Z",
     "iopub.status.busy": "2023-10-28T03:01:55.005415Z",
     "iopub.status.idle": "2023-10-28T03:01:55.011286Z",
     "shell.execute_reply": "2023-10-28T03:01:55.010549Z"
    },
    "id": "-QygoFvB0cs5",
    "outputId": "e60d81bd-8420-4cd7-e30f-3145f322bc20",
    "papermill": {
     "duration": 0.053387,
     "end_time": "2023-10-28T03:01:55.013166",
     "exception": false,
     "start_time": "2023-10-28T03:01:54.959779",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tensor([[1, 2, 3]])"
      ]
     },
     "execution_count": 86,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Select all elements from the first column (index 0) along the second dimension of the tensor x\n",
    "selected_column = x[:, 0]\n",
    "selected_column"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "id": "03e2c39f",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:55.106198Z",
     "iopub.status.busy": "2023-10-28T03:01:55.105902Z",
     "iopub.status.idle": "2023-10-28T03:01:55.111850Z",
     "shell.execute_reply": "2023-10-28T03:01:55.111028Z"
    },
    "id": "38Py4xWk0riy",
    "outputId": "268e5928-f9e6-44f6-dc17-c7de922e9401",
    "papermill": {
     "duration": 0.053801,
     "end_time": "2023-10-28T03:01:55.113716",
     "exception": false,
     "start_time": "2023-10-28T03:01:55.059915",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tensor([[2, 5, 8]])"
      ]
     },
     "execution_count": 87,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Get all values of 0th & 1st dimensions but only index 1 of 2nd dimension\n",
    "x[:, :, 1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 88,
   "id": "06c594d0",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:55.207761Z",
     "iopub.status.busy": "2023-10-28T03:01:55.207479Z",
     "iopub.status.idle": "2023-10-28T03:01:55.213538Z",
     "shell.execute_reply": "2023-10-28T03:01:55.212709Z"
    },
    "id": "5rfWF_D901My",
    "outputId": "50556de6-5c9a-41eb-f839-0c69234f13c4",
    "papermill": {
     "duration": 0.055551,
     "end_time": "2023-10-28T03:01:55.215553",
     "exception": false,
     "start_time": "2023-10-28T03:01:55.160002",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tensor([5])"
      ]
     },
     "execution_count": 88,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Get all values of the 0 dimension but only the 1 index value of the 1st and 2nd dimension\n",
    "x[:, 1, 1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 89,
   "id": "a543ca7b",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:55.314931Z",
     "iopub.status.busy": "2023-10-28T03:01:55.314605Z",
     "iopub.status.idle": "2023-10-28T03:01:55.321228Z",
     "shell.execute_reply": "2023-10-28T03:01:55.320365Z"
    },
    "id": "Nu5m-Sej1PHx",
    "outputId": "b4ad86b4-0234-4a6d-d3c9-c8ebf07ae892",
    "papermill": {
     "duration": 0.058472,
     "end_time": "2023-10-28T03:01:55.323224",
     "exception": false,
     "start_time": "2023-10-28T03:01:55.264752",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tensor([1, 2, 3])"
      ]
     },
     "execution_count": 89,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Get index 0 of 0th and 1st dimension and all values of 2nd dimension\n",
    "x[0, 0, :] # same as x[0][0]"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "414b506d",
   "metadata": {
    "id": "PPpz3D7B1RYR",
    "papermill": {
     "duration": 0.047769,
     "end_time": "2023-10-28T03:01:55.420347",
     "exception": false,
     "start_time": "2023-10-28T03:01:55.372578",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "\n",
    "## **Tensors & NumPy**\n",
    "PyTorch provides convenient methods for converting between NumPy arrays and PyTorch tensors:\n",
    "\n",
    "- **torch.from_numpy(ndarray)** converts a NumPy array to a PyTorch tensor.\n",
    "- **torch.Tensor.numpy()** converts a PyTorch tensor to a NumPy array.\n",
    "\n",
    "These methods allow for seamless interaction between NumPy and PyTorch data structures."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 90,
   "id": "1f3070d9",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:55.514798Z",
     "iopub.status.busy": "2023-10-28T03:01:55.514492Z",
     "iopub.status.idle": "2023-10-28T03:01:55.520919Z",
     "shell.execute_reply": "2023-10-28T03:01:55.519988Z"
    },
    "id": "RIJwa06Q2skP",
    "outputId": "576a53c5-d212-4153-fe2e-e0db49f21f31",
    "papermill": {
     "duration": 0.05568,
     "end_time": "2023-10-28T03:01:55.522972",
     "exception": false,
     "start_time": "2023-10-28T03:01:55.467292",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(array([1., 2., 3., 4., 5., 6., 7.]), (7,))"
      ]
     },
     "execution_count": 90,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Create a NumPy array with values ranging from 1.0 to 7.0\n",
    "array = np.arange(1.0, 8.0)\n",
    "\n",
    "# NumPy array and its shape\n",
    "array, array.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 91,
   "id": "8a5535f2",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:55.617401Z",
     "iopub.status.busy": "2023-10-28T03:01:55.616845Z",
     "iopub.status.idle": "2023-10-28T03:01:55.623683Z",
     "shell.execute_reply": "2023-10-28T03:01:55.622766Z"
    },
    "id": "leGXAwDD6kEl",
    "outputId": "8b3af961-5002-4b52-d35d-5470cf8172ef",
    "papermill": {
     "duration": 0.055388,
     "end_time": "2023-10-28T03:01:55.625807",
     "exception": false,
     "start_time": "2023-10-28T03:01:55.570419",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(tensor([1., 2., 3., 4., 5., 6., 7.], dtype=torch.float64), torch.Size([7]))"
      ]
     },
     "execution_count": 91,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Convert the NumPy array 'array' to a PyTorch tensor\n",
    "tensor = torch.from_numpy(array)\n",
    "\n",
    "# PyTorch tensor and its shape\n",
    "tensor, tensor.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 92,
   "id": "59ea2d58",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:55.720948Z",
     "iopub.status.busy": "2023-10-28T03:01:55.720665Z",
     "iopub.status.idle": "2023-10-28T03:01:55.727522Z",
     "shell.execute_reply": "2023-10-28T03:01:55.726533Z"
    },
    "id": "mTPieq1O6p38",
    "outputId": "affb531a-c584-4586-955b-cb3ccabdbdde",
    "papermill": {
     "duration": 0.055495,
     "end_time": "2023-10-28T03:01:55.729429",
     "exception": false,
     "start_time": "2023-10-28T03:01:55.673934",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(array([2., 3., 4., 5., 6., 7., 8.]),\n",
       " tensor([1., 2., 3., 4., 5., 6., 7.], dtype=torch.float64))"
      ]
     },
     "execution_count": 92,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Change the array, keep the tensor\n",
    "array = array + 1\n",
    "array, tensor"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 93,
   "id": "a6bea05b",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:55.825087Z",
     "iopub.status.busy": "2023-10-28T03:01:55.824501Z",
     "iopub.status.idle": "2023-10-28T03:01:55.832107Z",
     "shell.execute_reply": "2023-10-28T03:01:55.831280Z"
    },
    "id": "3X-8kmSe67tV",
    "outputId": "a0938d25-e90a-4ea4-cc1c-978051a8196f",
    "papermill": {
     "duration": 0.05631,
     "end_time": "2023-10-28T03:01:55.834099",
     "exception": false,
     "start_time": "2023-10-28T03:01:55.777789",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(tensor([1., 1., 1., 1., 1., 1., 1.]),\n",
       " array([1., 1., 1., 1., 1., 1., 1.], dtype=float32))"
      ]
     },
     "execution_count": 93,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Create a PyTorch tensor 'tensor' filled with ones and having dtype=float32\n",
    "tensor = torch.ones(7)\n",
    "\n",
    "# Convert the PyTorch tensor to a NumPy array 'numpy_tensor'\n",
    "numpy_tensor = tensor.numpy()\n",
    "\n",
    "# Display the PyTorch tensor and the NumPy array\n",
    "tensor, numpy_tensor"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f7733d10",
   "metadata": {
    "id": "WTaTjDHH7BiN",
    "papermill": {
     "duration": 0.047815,
     "end_time": "2023-10-28T03:01:55.929041",
     "exception": false,
     "start_time": "2023-10-28T03:01:55.881226",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "## **Reproducibility in PyTorch**\n",
    "Randomness is a powerful tool, but sometimes you need to reduce it for the sake of reproducibility. Reproducibility allows you and others to obtain the same or highly similar results when running the same code on different machines.\n",
    "\n",
    "Consider a scenario where you've developed an algorithm achieving a specific performance level, and you want your colleague to validate your findings on their computer. How can you ensure consistent results?\n",
    "\n",
    "This is where reproducibility comes into play. In PyTorch, you can take steps to control randomness, making your experiments more predictable.\n",
    "\n",
    "To demonstrate this, let's start by creating two random tensors. Since they are random, you might expect them to be different, right?\n",
    "\n",
    "Here's an example of ensuring reproducibility in PyTorch."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 94,
   "id": "f7ac14d5",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:56.029128Z",
     "iopub.status.busy": "2023-10-28T03:01:56.028398Z",
     "iopub.status.idle": "2023-10-28T03:01:56.039042Z",
     "shell.execute_reply": "2023-10-28T03:01:56.038202Z"
    },
    "id": "uS52jzmG8erb",
    "outputId": "f78d4b02-25ab-4497-ce01-4be59d289511",
    "papermill": {
     "duration": 0.062729,
     "end_time": "2023-10-28T03:01:56.041022",
     "exception": false,
     "start_time": "2023-10-28T03:01:55.978293",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Tensor A:\n",
      "tensor([[0.2741, 0.6142, 0.8973, 0.3629],\n",
      "        [0.1748, 0.2401, 0.5457, 0.7303],\n",
      "        [0.5268, 0.6694, 0.3213, 0.4008]])\n",
      "\n",
      "Tensor B:\n",
      "tensor([[0.2892, 0.9977, 0.6649, 0.5646],\n",
      "        [0.9323, 0.4621, 0.4027, 0.1680],\n",
      "        [0.1170, 0.5063, 0.6061, 0.5141]])\n",
      "\n",
      "Does Tensor A equal Tensor B? (anywhere)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "tensor([[False, False, False, False],\n",
       "        [False, False, False, False],\n",
       "        [False, False, False, False]])"
      ]
     },
     "execution_count": 94,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Create two random tensors, random_tensor_A and random_tensor_B\n",
    "random_tensor_A = torch.rand(3, 4)\n",
    "random_tensor_B = torch.rand(3, 4)\n",
    "\n",
    "# Display the values of Tensor A and Tensor B\n",
    "print(f\"Tensor A:\\n{random_tensor_A}\\n\")\n",
    "print(f\"Tensor B:\\n{random_tensor_B}\\n\")\n",
    "\n",
    "# Check if Tensor A equals Tensor B (element-wise, anywhere)\n",
    "are_equal = random_tensor_A == random_tensor_B\n",
    "\n",
    "# Print the result of the equality check\n",
    "print(f\"Does Tensor A equal Tensor B? (anywhere)\")\n",
    "are_equal"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "714a4522",
   "metadata": {
    "id": "Icz7oUcI8pRT",
    "papermill": {
     "duration": 0.048624,
     "end_time": "2023-10-28T03:01:56.139561",
     "exception": false,
     "start_time": "2023-10-28T03:01:56.090937",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "As expected, the previous random tensors had different values. But what if you wanted to create two random tensors with the same \"flavor\" of randomness? In other words, you want the tensors to still contain random values, but these values should be consistent.\n",
    "\n",
    "This is where torch.manual_seed(seed) comes into play. You can specify an integer value (such as 42 or any other) as the seed, and it will determine the characteristics of randomness.\n",
    "\n",
    "Let's demonstrate this by creating some random tensors with the same flavor."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "id": "d0c34761",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:56.239204Z",
     "iopub.status.busy": "2023-10-28T03:01:56.238413Z",
     "iopub.status.idle": "2023-10-28T03:01:56.247685Z",
     "shell.execute_reply": "2023-10-28T03:01:56.246797Z"
    },
    "id": "iPJXBEjB9JWa",
    "papermill": {
     "duration": 0.061967,
     "end_time": "2023-10-28T03:01:56.249852",
     "exception": false,
     "start_time": "2023-10-28T03:01:56.187885",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tensor([[0.8823, 0.9150, 0.3829, 0.9593],\n",
       "        [0.3904, 0.6009, 0.2566, 0.7936],\n",
       "        [0.9408, 0.1332, 0.9346, 0.5936]])"
      ]
     },
     "execution_count": 95,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Import the random module\n",
    "import random\n",
    "\n",
    "# Set the random seed to ensure reproducibility\n",
    "RANDOM_SEED = 42  # You can change this value to experiment with different random seeds\n",
    "\n",
    "# Set the random seed for PyTorch\n",
    "torch.manual_seed(seed=RANDOM_SEED)\n",
    "\n",
    "# Create a random tensor with the specified random seed\n",
    "random_tensor_C = torch.rand(3, 4)\n",
    "random_tensor_C"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 96,
   "id": "698a2dc4",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:56.349526Z",
     "iopub.status.busy": "2023-10-28T03:01:56.349189Z",
     "iopub.status.idle": "2023-10-28T03:01:56.354288Z",
     "shell.execute_reply": "2023-10-28T03:01:56.353596Z"
    },
    "id": "gK5Kk0Ie9K8C",
    "papermill": {
     "duration": 0.05651,
     "end_time": "2023-10-28T03:01:56.356223",
     "exception": false,
     "start_time": "2023-10-28T03:01:56.299713",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "# Have to reset the seed every time a new rand() is called\n",
    "# Without this, tensor_D would be different to tensor_C\n",
    "torch.random.manual_seed(seed=RANDOM_SEED) # try commenting this line out and seeing what happens\n",
    "random_tensor_D = torch.rand(3, 4)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 97,
   "id": "c6804d77",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:56.452430Z",
     "iopub.status.busy": "2023-10-28T03:01:56.452090Z",
     "iopub.status.idle": "2023-10-28T03:01:56.459940Z",
     "shell.execute_reply": "2023-10-28T03:01:56.459243Z"
    },
    "id": "0HaEs1Tl9dK0",
    "outputId": "9a740670-b11d-40c2-c4f2-3133d9710bf8",
    "papermill": {
     "duration": 0.057177,
     "end_time": "2023-10-28T03:01:56.461701",
     "exception": false,
     "start_time": "2023-10-28T03:01:56.404524",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Tensor C:\n",
      "tensor([[0.8823, 0.9150, 0.3829, 0.9593],\n",
      "        [0.3904, 0.6009, 0.2566, 0.7936],\n",
      "        [0.9408, 0.1332, 0.9346, 0.5936]])\n",
      "\n",
      "Tensor D:\n",
      "tensor([[0.8823, 0.9150, 0.3829, 0.9593],\n",
      "        [0.3904, 0.6009, 0.2566, 0.7936],\n",
      "        [0.9408, 0.1332, 0.9346, 0.5936]])\n",
      "\n",
      "Does Tensor C equal Tensor D? (anywhere)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "tensor([[True, True, True, True],\n",
       "        [True, True, True, True],\n",
       "        [True, True, True, True]])"
      ]
     },
     "execution_count": 97,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print(f\"Tensor C:\\n{random_tensor_C}\\n\")\n",
    "print(f\"Tensor D:\\n{random_tensor_D}\\n\")\n",
    "print(f\"Does Tensor C equal Tensor D? (anywhere)\")\n",
    "random_tensor_C == random_tensor_D"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9f110bcd",
   "metadata": {
    "id": "IFZifm1g9rsL",
    "papermill": {
     "duration": 0.047276,
     "end_time": "2023-10-28T03:01:56.558087",
     "exception": false,
     "start_time": "2023-10-28T03:01:56.510811",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "## **Device Compatibility**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 98,
   "id": "b3b38cce",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:56.654885Z",
     "iopub.status.busy": "2023-10-28T03:01:56.654108Z",
     "iopub.status.idle": "2023-10-28T03:01:57.710835Z",
     "shell.execute_reply": "2023-10-28T03:01:57.709885Z"
    },
    "id": "ANCTymNUHnKQ",
    "outputId": "4b5787d6-2a9e-44f8-af46-ada2d6c11796",
    "papermill": {
     "duration": 1.1082,
     "end_time": "2023-10-28T03:01:57.713051",
     "exception": false,
     "start_time": "2023-10-28T03:01:56.604851",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Sat Oct 28 03:01:57 2023       \r\n",
      "+-----------------------------------------------------------------------------+\r\n",
      "| NVIDIA-SMI 470.161.03   Driver Version: 470.161.03   CUDA Version: 11.4     |\r\n",
      "|-------------------------------+----------------------+----------------------+\r\n",
      "| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |\r\n",
      "| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |\r\n",
      "|                               |                      |               MIG M. |\r\n",
      "|===============================+======================+======================|\r\n",
      "|   0  Tesla T4            Off  | 00000000:00:04.0 Off |                    0 |\r\n",
      "| N/A   37C    P8    11W /  70W |      0MiB / 15109MiB |      0%      Default |\r\n",
      "|                               |                      |                  N/A |\r\n",
      "+-------------------------------+----------------------+----------------------+\r\n",
      "|   1  Tesla T4            Off  | 00000000:00:05.0 Off |                    0 |\r\n",
      "| N/A   38C    P8     9W /  70W |      0MiB / 15109MiB |      0%      Default |\r\n",
      "|                               |                      |                  N/A |\r\n",
      "+-------------------------------+----------------------+----------------------+\r\n",
      "                                                                               \r\n",
      "+-----------------------------------------------------------------------------+\r\n",
      "| Processes:                                                                  |\r\n",
      "|  GPU   GI   CI        PID   Type   Process name                  GPU Memory |\r\n",
      "|        ID   ID                                                   Usage      |\r\n",
      "|=============================================================================|\r\n",
      "|  No running processes found                                                 |\r\n",
      "+-----------------------------------------------------------------------------+\r\n"
     ]
    }
   ],
   "source": [
    "# This command, '!nvidia-smi', is used to display information about the NVIDIA GPU(s) installed on the system. \n",
    "# It provides details such as GPU model, memory usage, temperature, and more. \n",
    "# It's a useful tool for monitoring GPU status and can be helpful for tasks like deep learning training.\n",
    "!nvidia-smi"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0ce1b7af",
   "metadata": {
    "papermill": {
     "duration": 0.046192,
     "end_time": "2023-10-28T03:01:57.806371",
     "exception": false,
     "start_time": "2023-10-28T03:01:57.760179",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "#### **Checking Device**\n",
    "Checking the device (CPU or GPU) is essential for efficient utilization of hardware resources. It allows for faster computation on GPUs, effective memory management, and compatibility with deep learning libraries, ultimately improving performance in machine learning tasks."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 99,
   "id": "27764388",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:57.901662Z",
     "iopub.status.busy": "2023-10-28T03:01:57.901324Z",
     "iopub.status.idle": "2023-10-28T03:01:58.000872Z",
     "shell.execute_reply": "2023-10-28T03:01:57.999951Z"
    },
    "id": "oVAOswxgHnnJ",
    "outputId": "d9a1d793-2cb3-4649-a8f2-c2325a9ebb84",
    "papermill": {
     "duration": 0.149258,
     "end_time": "2023-10-28T03:01:58.002852",
     "exception": false,
     "start_time": "2023-10-28T03:01:57.853594",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "True"
      ]
     },
     "execution_count": 99,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Check if a GPU is available for use with PyTorch\n",
    "gpu_available = torch.cuda.is_available()\n",
    "\n",
    "# Print the result (True if GPU is available, False otherwise)\n",
    "gpu_available"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d0d260c6",
   "metadata": {
    "papermill": {
     "duration": 0.047557,
     "end_time": "2023-10-28T03:01:58.098601",
     "exception": false,
     "start_time": "2023-10-28T03:01:58.051044",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "#### **Setting Device**\n",
    "Setting a device in deep learning is the act of specifying whether computations should occur on a CPU or GPU. A \"cpu\" device implies that all operations, including model training and inference, are performed on the Central Processing Unit. In contrast, a \"cuda\" or GPU device offloads computations to the Graphics Processing Unit, which excels at parallel tasks like deep learning. The choice depends on hardware availability, model size, memory requirements, and compatibility with libraries, impacting training speed and resource utilization. It allows developers to optimize deep learning workflows for their specific hardware configuration."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 100,
   "id": "cb23bd62",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:58.194913Z",
     "iopub.status.busy": "2023-10-28T03:01:58.194236Z",
     "iopub.status.idle": "2023-10-28T03:01:58.200466Z",
     "shell.execute_reply": "2023-10-28T03:01:58.199595Z"
    },
    "id": "V-PTwT6RHxAJ",
    "outputId": "5b7ff4d0-4778-431f-c3f9-7330f7ec14d2",
    "papermill": {
     "duration": 0.057163,
     "end_time": "2023-10-28T03:01:58.202450",
     "exception": false,
     "start_time": "2023-10-28T03:01:58.145287",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'cuda'"
      ]
     },
     "execution_count": 100,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Check if a GPU is available, and if so, set the device type to \"cuda\"; otherwise, use \"cpu\"\n",
    "device = \"cuda\" if torch.cuda.is_available() else \"cpu\"\n",
    "\n",
    "# Print the selected device type\n",
    "device"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 101,
   "id": "ec6d520e",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:58.302789Z",
     "iopub.status.busy": "2023-10-28T03:01:58.302035Z",
     "iopub.status.idle": "2023-10-28T03:01:58.328017Z",
     "shell.execute_reply": "2023-10-28T03:01:58.327167Z"
    },
    "id": "52I_Y774IL06",
    "outputId": "d114f6c7-4e4e-45e4-dd84-50a4117d59ed",
    "papermill": {
     "duration": 0.078393,
     "end_time": "2023-10-28T03:01:58.330044",
     "exception": false,
     "start_time": "2023-10-28T03:01:58.251651",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "2"
      ]
     },
     "execution_count": 101,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Count the number of devices\n",
    "torch.cuda.device_count()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9a8f178c",
   "metadata": {
    "id": "g6XuQu6_IPSb",
    "papermill": {
     "duration": 0.047188,
     "end_time": "2023-10-28T03:01:58.425015",
     "exception": false,
     "start_time": "2023-10-28T03:01:58.377827",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "#### **Putting tensors (and models) on the GPU**\n",
    "**Note:** Putting a tensor on GPU using to(device) (e.g. some_tensor.to(device)) returns a copy of that tensor, e.g. the same tensor will be on CPU and GPU. To overwrite tensors, reassign them:\n",
    "\n",
    "some_tensor = some_tensor.to(device)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 102,
   "id": "67291a51",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:01:58.519640Z",
     "iopub.status.busy": "2023-10-28T03:01:58.519327Z",
     "iopub.status.idle": "2023-10-28T03:02:01.697658Z",
     "shell.execute_reply": "2023-10-28T03:02:01.696732Z"
    },
    "id": "ZMmH3xtGISoZ",
    "outputId": "d7d9e018-55eb-4cbe-cf2e-e63f762050a6",
    "papermill": {
     "duration": 3.228459,
     "end_time": "2023-10-28T03:02:01.699741",
     "exception": false,
     "start_time": "2023-10-28T03:01:58.471282",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "tensor([1, 2, 3]) cpu\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "tensor([1, 2, 3], device='cuda:0')"
      ]
     },
     "execution_count": 102,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Create tensor (default on CPU)\n",
    "tensor = torch.tensor([1, 2, 3])\n",
    "\n",
    "# Tensor not on GPU\n",
    "print(tensor, tensor.device)\n",
    "\n",
    "# Move tensor to GPU (if available)\n",
    "tensor_on_gpu = tensor.to(device)\n",
    "tensor_on_gpu"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a906e871",
   "metadata": {
    "id": "sB2B2H3hIjo6",
    "papermill": {
     "duration": 0.04726,
     "end_time": "2023-10-28T03:02:01.795393",
     "exception": false,
     "start_time": "2023-10-28T03:02:01.748133",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "#### **Moving tensors back to the CPU**"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "36869755",
   "metadata": {
    "papermill": {
     "duration": 0.047006,
     "end_time": "2023-10-28T03:02:01.888920",
     "exception": false,
     "start_time": "2023-10-28T03:02:01.841914",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "**If tensor is on GPU, can't transform it to NumPy (this will error)**\n",
    "\n",
    "- For example, as the tensor tensor_on_gpu is on GPU now, so if we run the given code then it will show the error as follow:\n",
    "\n",
    "> ***tensor_on_gpu.numpy()***"
   ]
  },
  {
   "attachments": {
    "c89d7b36-3878-473c-a255-b540311ae6bb.png": {
     "image/png": "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"
    }
   },
   "cell_type": "markdown",
   "id": "7d364aca",
   "metadata": {
    "papermill": {
     "duration": 0.046448,
     "end_time": "2023-10-28T03:02:01.982119",
     "exception": false,
     "start_time": "2023-10-28T03:02:01.935671",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "![image.png](attachment:c89d7b36-3878-473c-a255-b540311ae6bb.png)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 103,
   "id": "f92822c4",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:02:02.080933Z",
     "iopub.status.busy": "2023-10-28T03:02:02.079992Z",
     "iopub.status.idle": "2023-10-28T03:02:02.087132Z",
     "shell.execute_reply": "2023-10-28T03:02:02.086227Z"
    },
    "id": "cY20BL-3I1YB",
    "outputId": "cd7edb3e-c0b4-4050-d385-e5217aea31cd",
    "papermill": {
     "duration": 0.059995,
     "end_time": "2023-10-28T03:02:02.089344",
     "exception": false,
     "start_time": "2023-10-28T03:02:02.029349",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([1, 2, 3])"
      ]
     },
     "execution_count": 103,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Instead, copy the tensor back to cpu\n",
    "tensor_back_on_cpu = tensor_on_gpu.cpu().numpy()\n",
    "tensor_back_on_cpu"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 104,
   "id": "de48c649",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2023-10-28T03:02:02.199222Z",
     "iopub.status.busy": "2023-10-28T03:02:02.198612Z",
     "iopub.status.idle": "2023-10-28T03:02:02.206122Z",
     "shell.execute_reply": "2023-10-28T03:02:02.204978Z"
    },
    "id": "xBRSbfS1I4dC",
    "outputId": "761084bb-41d9-4c70-bc0d-2f6df2c4eb66",
    "papermill": {
     "duration": 0.06919,
     "end_time": "2023-10-28T03:02:02.208741",
     "exception": false,
     "start_time": "2023-10-28T03:02:02.139551",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tensor([1, 2, 3], device='cuda:0')"
      ]
     },
     "execution_count": 104,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tensor_on_gpu"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "83c742f7",
   "metadata": {
    "papermill": {
     "duration": 0.060604,
     "end_time": "2023-10-28T03:02:02.325165",
     "exception": false,
     "start_time": "2023-10-28T03:02:02.264561",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "## **Closing Thoughts**\n",
    "In this comprehensive guide, we've explored the fundamental concepts of PyTorch, a powerful deep learning framework, and its practical applications. We started by introducing PyTorch and its essential libraries, followed by a hands-on journey through tensors, their operations, and various tensor types. We discussed GPU utilization for faster computations and how to ensure reproducibility in your experiments. Whether you're a beginner in deep learning or an experienced practitioner, this guide serves as a valuable resource for harnessing the capabilities of PyTorch in your machine learning projects. With its flexibility, performance optimization, and rich ecosystem, PyTorch continues to be a driving force in the world of deep learning. Now armed with this knowledge, you're well-equipped to dive deeper into PyTorch and explore its vast potential in your data science endeavors.\n",
    "\n",
    "**Please feel free to ask in the comment section if you have any confusion or questions.**"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "dd433676",
   "metadata": {
    "papermill": {
     "duration": 0.054522,
     "end_time": "2023-10-28T03:02:02.430541",
     "exception": false,
     "start_time": "2023-10-28T03:02:02.376019",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "#### Here are some of the contributions I've made on Kaggle:\n",
    "\n",
    "- **[Comprehensive Guide on NumPy for Beginners](https://www.kaggle.com/code/tanvirnwu/comprehensive-guide-on-numpy-for-beginners#Learn-More)**\n",
    "- **[Comprehensive Guide on NumPy for Beginners](https://www.kaggle.com/code/tanvirnwu/boolean-in-python-with-example-and-explanation)**\n",
    "- **[Dictionary in Python with Examples & Explanations](https://www.kaggle.com/code/tanvirnwu/dictionary-in-python-with-examples-explanations)**\n",
    "- **[List in Python for Beginners](https://www.kaggle.com/code/tanvirnwu/list-in-python-for-beginners)**\n",
    "- **[A Brief Introduction of Graph Neural Network (GNN): Concepts, Types, and Uses](https://www.kaggle.com/discussions/general/449125#2493256)**\n",
    "- **[Essential Python Libraries for Data Visualization](https://www.kaggle.com/discussions/getting-started/450857)**"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "affbee14",
   "metadata": {
    "papermill": {
     "duration": 0.047442,
     "end_time": "2023-10-28T03:02:02.526012",
     "exception": false,
     "start_time": "2023-10-28T03:02:02.478570",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "### **Please Inspire me by your Upvotes, Comment & Follow. Thanks!**"
   ]
  }
 ],
 "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": "3.10.12"
  },
  "papermill": {
   "default_parameters": {},
   "duration": 29.877621,
   "end_time": "2023-10-28T03:02:04.936280",
   "environment_variables": {},
   "exception": null,
   "input_path": "__notebook__.ipynb",
   "output_path": "__notebook__.ipynb",
   "parameters": {},
   "start_time": "2023-10-28T03:01:35.058659",
   "version": "2.4.0"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}