【pytorch官方文档学习之二】tensors

本系列旨在通过阅读官方pytorch代码熟悉CNN各个框架的实现方式和流程。

【pytorch官方文档学习之二】tensors

本文是对官方文档PyTorch: Tensors的详细注释和个人理解，欢迎交流。

• tensor 与 numpy的区别
  Behind the scenes, Tensors can keep track of a computational graph and gradients, but they’re also useful as a generic tool for scientific computing. Also unlike numpy, PyTorch Tensors can utilize GPUs to accelerate their numeric computations. To run a PyTorch Tensor on GPU, you simply need to cast it to a new datatype.
• 实例

# -*- coding: utf-8 -*-

import torch


dtype = torch.float
device = torch.device("cpu")
# device = torch.device("cuda:0") # Uncomment this to run on GPU

# N is batch size; D_in is input dimension;
# H is hidden dimension; D_out is output dimension.
N, D_in, H, D_out = 64, 1000, 100, 10

# Create random input and output data
x = torch.randn(N, D_in, device=device, dtype=dtype) # 定义输入数据及维度并传入cuda
y = torch.randn(N, D_out, device=device, dtype=dtype) # 定义输出数据及维度并传入cuda

# Randomly initialize weights
w1 = torch.randn(D_in, H, device=device, dtype=dtype) # 随机初始化weights并传入cuda
w2 = torch.randn(H, D_out, device=device, dtype=dtype)

learning_rate = 1e-6   #定义lr
for t in range(500):
    # Forward pass: compute predicted y   forward
    h = x.mm(w1)
    h_relu = h.clamp(min=0)  # torch.clamp(input, min, max, out = None)→tensor 将输入input张量每个元素的夹紧到区间 [min,max]，并返回结果到一个新张量。relu的又一实现方法。
    y_pred = h_relu.mm(w2)

    # Compute and print loss
    loss = (y_pred - y).pow(2).sum().item()
    if t % 100 == 99:    # 迭代99次
        print(t, loss)

    # Backprop to compute gradients of w1 and w2 with respect to loss   反向传播 
    grad_y_pred = 2.0 * (y_pred - y)
    grad_w2 = h_relu.t().mm(grad_y_pred)
    grad_h_relu = grad_y_pred.mm(w2.t())
    grad_h = grad_h_relu.clone()
    grad_h[h < 0] = 0
    grad_w1 = x.t().mm(grad_h)

    # Update weights using gradient descent   更新weights
    w1 -= learning_rate * grad_w1
    w2 -= learning_rate * grad_w2

除cuda内容之外，tensor的框架跟numpy一致。