{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Tensor 张量\n",
    "Tensor类似于Numpy的ndarray，但不同的是，Tensor可以在GPU上被加速。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "from __future__ import print_function\n",
    "import torch"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "生成一个未初始化的 5×3 矩阵："
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "tensor([[0., 0., 0.],\n",
      "        [0., 0., 0.],\n",
      "        [0., 0., 0.],\n",
      "        [0., 0., 0.],\n",
      "        [0., 0., 0.]])\n"
     ]
    }
   ],
   "source": [
    "x = torch.empty(5,3)\n",
    "print(x)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "生成一个随机矩阵："
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "tensor([[0.4907, 0.1470, 0.9752],\n",
      "        [0.9518, 0.6973, 0.4775],\n",
      "        [0.5643, 0.6586, 0.3142],\n",
      "        [0.2364, 0.8435, 0.6187],\n",
      "        [0.3253, 0.5903, 0.9939]])\n"
     ]
    }
   ],
   "source": [
    "x = torch.rand(5,3)\n",
    "print(x)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "生成一个0矩阵，数据类型为long"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "tensor([[0, 0, 0],\n",
      "        [0, 0, 0],\n",
      "        [0, 0, 0],\n",
      "        [0, 0, 0],\n",
      "        [0, 0, 0]])\n"
     ]
    }
   ],
   "source": [
    "x = torch.zeros(5,3,dtype=torch.long)\n",
    "print(x)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "直接从数据创造tensor"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "tensor([6.7000, 6.0000, 7.0000])\n"
     ]
    }
   ],
   "source": [
    "x = torch.tensor([6.7,6,7])\n",
    "print(x)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "在已有tensor的基础上创建tensor。这些方法将会继承原tensor的属性，比如数据类型，除非新的值由用户给出。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "x:\n",
      " tensor([6.7000, 6.0000, 7.0000])\n",
      "y:\n",
      " tensor([[1., 1., 1.],\n",
      "        [1., 1., 1.],\n",
      "        [1., 1., 1.],\n",
      "        [1., 1., 1.],\n",
      "        [1., 1., 1.]], dtype=torch.float64)\n",
      "tensor([-0.3186,  0.5506,  0.4940])\n"
     ]
    }
   ],
   "source": [
    "print('x:\\n',x)\n",
    "y = x.new_ones(5, 3, dtype=torch.double)\n",
    "print('y:\\n',y)\n",
    "\n",
    "x = torch.randn_like(x, dtype=torch.float)    # dtype被改写，但是结果继承了size。\n",
    "print(x)                                      "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "获得tensor的size"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "torch.Size([3])\n"
     ]
    }
   ],
   "source": [
    "print(x.size())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Operations 操作\n",
    "## 相加\n",
    "PyTorch中有许多种相加语法："
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**相加：**语法1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "tensor([[0.9943, 0.3976, 0.6645],\n",
      "        [0.6796, 1.1703, 1.2389],\n",
      "        [1.5879, 1.1208, 1.5912],\n",
      "        [1.2165, 0.8195, 1.0364],\n",
      "        [1.2178, 1.3195, 1.1736]])\n"
     ]
    }
   ],
   "source": [
    "x = torch.rand(5, 3)\n",
    "y = torch.rand(5, 3)\n",
    "print(x + y)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**相加：**语法2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "tensor([[0.9943, 0.3976, 0.6645],\n",
      "        [0.6796, 1.1703, 1.2389],\n",
      "        [1.5879, 1.1208, 1.5912],\n",
      "        [1.2165, 0.8195, 1.0364],\n",
      "        [1.2178, 1.3195, 1.1736]])\n"
     ]
    }
   ],
   "source": [
    "print(torch.add(x,y))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**相加：**将一个tensor作为out参数，用以输出。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "tensor([[0.9943, 0.3976, 0.6645],\n",
      "        [0.6796, 1.1703, 1.2389],\n",
      "        [1.5879, 1.1208, 1.5912],\n",
      "        [1.2165, 0.8195, 1.0364],\n",
      "        [1.2178, 1.3195, 1.1736]])\n"
     ]
    }
   ],
   "source": [
    "result = torch.empty(5,3)\n",
    "torch.add(x,y,out=result)\n",
    "print(result)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**相加：**替代"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "tensor([[0.9943, 0.3976, 0.6645],\n",
      "        [0.6796, 1.1703, 1.2389],\n",
      "        [1.5879, 1.1208, 1.5912],\n",
      "        [1.2165, 0.8195, 1.0364],\n",
      "        [1.2178, 1.3195, 1.1736]])\n"
     ]
    }
   ],
   "source": [
    "y.add_(x)\n",
    "print(y)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Note:\n",
    "> 任何让tensor被覆盖的方法的名称都以`_`结尾。例如`x.copy_(y)`, `x.t_()`等。"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 索引\n",
    "**NumPy中的索引方法可以使用：**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "tensor([0.0200, 0.4981, 0.2466, 0.7497, 0.8315])\n"
     ]
    }
   ],
   "source": [
    "print(x[:, 1])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Resizing  \n",
    "想要改变tensor的规模，可以使用`torch.view`。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "torch.Size([4, 4]) torch.Size([16]) torch.Size([2, 8])\n"
     ]
    }
   ],
   "source": [
    "x = torch.randn(4, 4)\n",
    "y = x.view(16)\n",
    "z = x.view(-1, 8)  # the size -1 is inferred from other dimensions\n",
    "print(x.size(), y.size(), z.size())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 取值\n",
    "如果你有一个单元素的tensor，`.item()`方法可以取值。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "tensor([-0.7093])\n",
      "-0.709314227104187\n"
     ]
    }
   ],
   "source": [
    "x = torch.randn(1)\n",
    "print(x)\n",
    "print(x.item())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 与NumPy的沟通桥梁\n",
    "将Torch张量转换为NumPy数组很容易，反之亦然。\n",
    "Torch张量将会和NumPy数组共享同一个内存位置（Torch张量在CPU上的情况下），其中一个发生改变，另一个也会随之改变。\n",
    "## 将Torch Tensor转换为NumPy数组"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "tensor([1., 1., 1., 1., 1.])\n",
      "[1. 1. 1. 1. 1.]\n",
      "tensor([2., 2., 2., 2., 2.])\n",
      "[2. 2. 2. 2. 2.]\n"
     ]
    }
   ],
   "source": [
    "a = torch.ones(5)\n",
    "print(a)\n",
    "b = a.numpy()\n",
    "print(b)\n",
    "a.add_(1)\n",
    "print(a)\n",
    "print(b)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 将NumPy数组转换为Torch Tensor"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[2. 2. 2. 2. 2.]\n",
      "tensor([2., 2., 2., 2., 2.], dtype=torch.float64)\n"
     ]
    }
   ],
   "source": [
    "import numpy as np\n",
    "a = np.ones(5)\n",
    "b = torch.from_numpy(a)\n",
    "np.add(a, 1, out=a)\n",
    "print(a)\n",
    "print(b)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "除了CharTensor，其他在CPU上的Torch张量均支持与NumPy数组的相互转换。"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# CUDA Tensors\n",
    "使用`.to()`方法，可以将张量转移到其他设备。"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "tensor([0.2907], device='cuda:0')\n",
      "tensor([0.2907], dtype=torch.float64)\n"
     ]
    }
   ],
   "source": [
    "if torch.cuda.is_available():\n",
    "    device = torch.device(\"cuda\")          # 创建CUDA设备对象\n",
    "    y = torch.ones_like(x, device=device)  # 直接在GPU上\n",
    "    x = x.to(device)\n",
    "    # x = x.to(\"cuda\") 也可以这么写\n",
    "    z = x + y\n",
    "    print(z)\n",
    "    print(z.to(\"cpu\", torch.double))       # ``.to`` 方法可以同时转换数据类型"
   ]
  }
 ],
 "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": "3.7.0"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}