{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 201 Torch and Numpy\n",
    "\n",
    "View more, visit my tutorial page: https://mofanpy.com/tutorials/\n",
    "My Youtube Channel: https://www.youtube.com/user/MorvanZhou\n",
    "\n",
    "Dependencies:\n",
    "* torch: 0.1.11\n",
    "* numpy\n",
    "\n",
    "Details about math operation in torch can be found in: http://pytorch.org/docs/torch.html#math-operations\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import torch\n",
    "import numpy as np"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\nnumpy array: [[0 1 2]\n [3 4 5]] \ntorch tensor: tensor([[ 0,  1,  2],\n        [ 3,  4,  5]], dtype=torch.int32) \ntensor to array: [[0 1 2]\n [3 4 5]]\n"
     ]
    }
   ],
   "source": [
    "# convert numpy to tensor or vise versa\n",
    "np_data = np.arange(6).reshape((2, 3))\n",
    "torch_data = torch.from_numpy(np_data)\n",
    "tensor2array = torch_data.numpy()\n",
    "print(\n",
    "    '\\nnumpy array:', np_data,          # [[0 1 2], [3 4 5]]\n",
    "    '\\ntorch tensor:', torch_data,      #  0  1  2 \\n 3  4  5    [torch.LongTensor of size 2x3]\n",
    "    '\\ntensor to array:', tensor2array, # [[0 1 2], [3 4 5]]\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\nabs \nnumpy:  [1 2 1 2] \ntorch:  tensor([ 1.,  2.,  1.,  2.])\n"
     ]
    }
   ],
   "source": [
    "# abs\n",
    "data = [-1, -2, 1, 2]\n",
    "tensor = torch.FloatTensor(data)  # 32-bit floating point\n",
    "print(\n",
    "    '\\nabs',\n",
    "    '\\nnumpy: ', np.abs(data),          # [1 2 1 2]\n",
    "    '\\ntorch: ', torch.abs(tensor)      # [1 2 1 2]\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tensor([ 1.,  2.,  1.,  2.])"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tensor.abs()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\nsin \nnumpy:  [-0.84147098 -0.90929743  0.84147098  0.90929743] \ntorch:  tensor([-0.8415, -0.9093,  0.8415,  0.9093])\n"
     ]
    }
   ],
   "source": [
    "# sin\n",
    "print(\n",
    "    '\\nsin',\n",
    "    '\\nnumpy: ', np.sin(data),      # [-0.84147098 -0.90929743  0.84147098  0.90929743]\n",
    "    '\\ntorch: ', torch.sin(tensor)  # [-0.8415 -0.9093  0.8415  0.9093]\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tensor([ 0.2689,  0.1192,  0.7311,  0.8808])"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tensor.sigmoid()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tensor([ 0.3679,  0.1353,  2.7183,  7.3891])"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tensor.exp()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\nmean \nnumpy:  0.0 \ntorch:  tensor(0.)\n"
     ]
    }
   ],
   "source": [
    "# mean\n",
    "print(\n",
    "    '\\nmean',\n",
    "    '\\nnumpy: ', np.mean(data),         # 0.0\n",
    "    '\\ntorch: ', torch.mean(tensor)     # 0.0\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\nmatrix multiplication (matmul) \nnumpy:  [[ 7 10]\n [15 22]] \ntorch:  tensor([[ 7., 10.],\n        [15., 22.]])\n"
     ]
    }
   ],
   "source": [
    "# matrix multiplication\n",
    "data = [[1,2], [3,4]]\n",
    "tensor = torch.FloatTensor(data)  # 32-bit floating point\n",
    "# correct method\n",
    "print(\n",
    "    '\\nmatrix multiplication (matmul)',\n",
    "    '\\nnumpy: ', np.matmul(data, data),     # [[7, 10], [15, 22]]\n",
    "    '\\ntorch: ', torch.mm(tensor, tensor)   # [[7, 10], [15, 22]]\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "ename": "RuntimeError",
     "evalue": "dot: Expected 1-D argument self, but got 2-D",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mRuntimeError\u001b[0m                              Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-3-a29f9258176b>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[1;32m      5\u001b[0m     \u001b[0;34m'\\nmatrix multiplication (dot)'\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      6\u001b[0m     \u001b[0;34m'\\nnumpy: '\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdata\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdot\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdata\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m        \u001b[0;31m# [[7, 10], [15, 22]]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 7\u001b[0;31m     \u001b[0;34m'\\ntorch: '\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtorch\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdot\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtensor\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdot\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtensor\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m     \u001b[0;31m# 30.0. Beware that torch.dot does not broadcast, only works for 1-dimensional tensor\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      8\u001b[0m )\n",
      "\u001b[0;31mRuntimeError\u001b[0m: dot: Expected 1-D argument self, but got 2-D"
     ],
     "output_type": "error"
    }
   ],
   "source": [
    "# incorrect method\n",
    "data = np.array(data)\n",
    "tensor = torch.Tensor(data)\n",
    "print(\n",
    "    '\\nmatrix multiplication (dot)',\n",
    "    '\\nnumpy: ', data.dot(data),        # [[7, 10], [15, 22]]\n",
    "    '\\ntorch: ', torch.dot(tensor.dot(tensor))     # NOT WORKING! Beware that torch.dot does not broadcast, only works for 1-dimensional tensor\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Note that:\n",
    "\n",
    "torch.dot(tensor1, tensor2) → float\n",
    "\n",
    "Computes the dot product (inner product) of two tensors. Both tensors are treated as 1-D vectors."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tensor([[  7.,  10.],\n        [ 15.,  22.]])"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tensor.mm(tensor)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tensor([[  1.,   4.],\n        [  9.,  16.]])"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tensor * tensor"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tensor(7.)"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "torch.dot(torch.Tensor([2, 3]), torch.Tensor([2, 1]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "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": "3.5.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}