{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Auto grad\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch\n",
    "import torchvision"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "tensor([5., 7., 9.], grad_fn=<AddBackward0>)\n",
      "<AddBackward0 object at 0x11f573ad0>\n",
      "tensor(21., grad_fn=<SumBackward0>)\n",
      "<SumBackward0 object at 0x11f595650>\n"
     ]
    }
   ],
   "source": [
    "# The tensor object keeps track of how it was created if requieres_grad is True \n",
    "x = torch.tensor([1.,2.,3],requires_grad=True)\n",
    "y = torch.tensor([4.,5.,6],requires_grad=True)\n",
    "\n",
    "z = x + y\n",
    "print(z)\n",
    "\n",
    "print(z.grad_fn)\n",
    "s = z.sum()\n",
    "\n",
    "print(s)\n",
    "\n",
    "print(s.grad_fn)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "tensor([2., 2., 2.])\n"
     ]
    }
   ],
   "source": [
    "# To back propagate\n",
    "s.backward()\n",
    "print(x.grad)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "False False\n",
      "None\n",
      "<AddBackward0 object at 0x11f088650>\n",
      "True\n",
      "None\n",
      "True\n",
      "True\n",
      "False\n"
     ]
    }
   ],
   "source": [
    "x = torch.randn(2,2)\n",
    "y = torch.randn(2,2)\n",
    "print(x.requires_grad,y.requires_grad)\n",
    "\n",
    "z = x + y\n",
    "\n",
    "print(z.grad_fn)\n",
    "\n",
    "x.requires_grad_()\n",
    "y.requires_grad_()\n",
    "\n",
    "z= x + y\n",
    "\n",
    "print(z.grad_fn)\n",
    "\n",
    "print(z.requires_grad)\n",
    "\n",
    "new_z = z.detach()\n",
    "\n",
    "print(new_z.grad_fn)\n",
    "\n",
    "print(x.requires_grad)\n",
    "print((x+10).requires_grad)\n",
    "\n",
    "with torch.no_grad():\n",
    "    print((x+10).requires_grad)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "tensor([[1., 1.],\n",
      "        [1., 1.]], requires_grad=True)\n",
      "tensor([[3., 3.],\n",
      "        [3., 3.]], grad_fn=<AddBackward0>)\n",
      "<AddBackward0 object at 0x11ee83b90>\n",
      "tensor([[27., 27.],\n",
      "        [27., 27.]], grad_fn=<MulBackward0>) tensor(27., grad_fn=<MeanBackward0>)\n",
      "tensor([[4.5000, 4.5000],\n",
      "        [4.5000, 4.5000]])\n"
     ]
    }
   ],
   "source": [
    "x = torch.ones(2,2,requires_grad=True)\n",
    "print(x)\n",
    "y = x + 2\n",
    "print(y)\n",
    "print(y.grad_fn)\n",
    "\n",
    "z = y*y*3\n",
    "\n",
    "out = z.mean()\n",
    "\n",
    "print(z,out)\n",
    "\n",
    "out.backward()\n",
    "print(x.grad)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}