python-3.x - 如何包装 PyTorch 函数并实现 autograd？

Question

我正在学习 Defining new autograd functions 上的 PyTorch 教程.我要实现的 autograd 函数是 torch.nn.functional.max_pool1d 的包装器。 .这是我到目前为止所拥有的:

import numpy as np
import torch
import torch.nn as nn
import torch.nn.functional as F
import torch.autograd as tag

class SquareAndMaxPool1d(tag.Function):

    @staticmethod
    def forward(ctx, input, kernel_size, stride=None, padding=0, dilation=1, \
                return_indices=False, ceil_mode=False):
        ctx.save_for_backward( input )

        inputC = input.clone() #copy input
        inputC *= inputC

        output = F.max_pool1d(inputC, kernel_size, stride=stride, \
                              padding=padding, dilation=dilation, \
                              return_indices=return_indices, \
                              ceil_mode=ceil_mode)

        return output

    @staticmethod
    def backward(ctx, grad_output):
        input, = ctx.saved_tensors
        grad_input = get_max_pool1d_grad_somehow(grad_output)
        return 2.0*input*grad_input

我的问题是:如何获得包装函数的渐变？我知道考虑到我提供的示例非常简单，可能还有其他方法可以做到这一点，但我想做的适合这个框架并且需要我实现 autograd功能。

编辑:检查后 this blog post我决定尝试以下 backward :

def backward(ctx, grad_output):
    input, output = ctx.saved_tensors
    grad_input = output.backward(grad_output)
    return 2.0*input*grad_input

与 output添加到已保存的变量中。然后我运行以下代码:

x = np.random.randn(1,1,5)
xT = torch.from_numpy(x)
xT.requires_grad=True
f = SquareAndMaxPool1d.apply
s = torch.sum(f(xT,2))
s.backward()

我得到 Bus error: 10 .

说， xT是 tensor([[[ 1.69533562, -0.21779421, 2.28693953, -0.86688095, -1.01033497]]], dtype=torch.float64) ，那么我希望找到 xT.grad是 tensor([[[ 3.39067124, -0. , 9.14775812, -0. , -2.02066994]]], dtype=torch.float64)调用 s.backward() 后(即 2*x*grad_of_max_pool ，其中 grad_of_max_pool 包含 tensor([[[1., 0., 2., 0., 1.]]], dtype=torch.float64) )。

我已经弄清楚为什么我会收到 Bus error: 10 .上面的代码似乎导致了我的 backward 的递归调用。在 grad_input = output.backward(grad_output) .所以我需要找到其他方法来获得 max_pool1d 的梯度.我知道如何在纯 Python 中实现它，但结果会比我可以包装库代码要慢得多。

Answer 1

你选择了一个相当不幸的例子。 torch.nn.functional.max_pool1d不是 torch.autograd.Function 的实例，因为它是 PyTorch 内置的，在 C++ 代码中定义并带有 autogenerated Python 绑定(bind)。我不确定是否有可能获得 backward属性通过其接口(interface)。

首先，如果你没有注意到，你不需要为这个公式的反向传播编写任何自定义代码，因为电源操作和 max_pool1d已经定义了它，因此它们的组成也包含在 autograd 中。假设您的目标是一项练习，我建议您更多地手动进行(不要退回到 backward 的 max_pool1d )。下面是一个例子

import torch
import torch.nn.functional as F
import torch.autograd as tag

class SquareAndMaxPool1d(tag.Function):
    @staticmethod
    def forward(ctx, input, kernel_size, **kwargs):
        # we're gonna need indices for backward. Currently SquareAnd...
        # never actually returns indices, I left it out for simplicity
        kwargs['return_indices'] = True

        input_sqr = input ** 2
        output, indices = F.max_pool1d(input_sqr, kernel_size, **kwargs)
        ctx.save_for_backward(input, indices)

        return output

    @staticmethod
    def backward(ctx, grad_output):
        input, indices = ctx.saved_tensors

        # first we need to reconstruct the gradient of `max_pool1d`
        # by putting all the output gradient elements (corresponding to
        # input elements which made it through the max_pool1d) in their
        # respective places, the rest has gradient of 0. We do it by
        # scattering it against a tensor of 0s
        grad_output_unpooled = torch.zeros_like(input)
        grad_output_unpooled.scatter_(2, indices, grad_output)

        # then incorporate the gradient of the "square" part of your
        # operator
        grad_input = 2. * input * grad_output_unpooled

        # the docs for backward
        # https://pytorch.org/docs/stable/autograd.html#torch.autograd.Function.backward
        # say that "it should return as many tensors, as there were inputs
        # to forward()". It fails to mention that if an argument was not a
        # tensor, it should return None (I remember reading this somewhere,
        # but can't find it anymore). Anyway, we need to
        # return a (grad_input, None) tuple to avoid a complaint that two
        # outputs were expected
        return grad_input, None

然后我们可以使用 numerical gradient checker验证操作是否按预期工作。

f = SquareAndMaxPool1d.apply
xT = torch.randn(1, 1, 6, requires_grad=True, dtype=torch.float64)
tag.gradcheck(lambda t: f(t, 2), xT)

如果这不能解决您关于如何获得 backward 的问题，我很抱歉。的 max_pool1d ，但希望您发现我的回答足够有用。