在 PyTorch 中使用阈值进行训练

Question

我有一个神经网络，它在输入时产生单一值。我需要使用网络返回的这个值来阈值另一个数组。该阈值操作的结果用于计算损失函数（阈值的值事先未知，需要通过训练得出）。以下是一个 MWE

import torch

x = torch.randn(10, 1)  # Say this is the output of the network (10 is my batch size)
data_array = torch.randn(10, 2)  # This is the data I need to threshold
ground_truth = torch.randn(10, 2)  # This is the ground truth
mse_loss = torch.nn.MSELoss()  # Loss function

# Threshold
thresholded_vals = data_array * (data_array >= x)  # Returns zero in all places where the value is less than the threshold, the value itself otherwise

# Compute loss and gradients
loss = mse_loss(thresholded_vals, ground_truth)
loss.backward()  # Throws error here

由于阈值操作 returns 一个没有任何梯度的张量数组 backward() 操作抛出错误。

在这种情况下如何训练网络？

Answer 1

您的阈值函数在阈值中不可微分，因此 torch 不会计算阈值的梯度，这就是您的示例不起作用的原因。

import torch

x = torch.randn(10, 1, requires_grad=True)  # Say this is the output of the network (10 is my batch size)
data_array = torch.randn(10, 2, requires_grad=True)  # This is the data I need to threshold
ground_truth = torch.randn(10, 2)  # This is the ground truth
mse_loss = torch.nn.MSELoss()  # Loss function

# Threshold
thresholded_vals = data_array * (data_array >= x)  # Returns zero in all places where the value is less than the threshold, the value itself otherwise

# Compute loss and gradients
loss = mse_loss(thresholded_vals, ground_truth)
loss.backward()  # Throws error here
print(x.grad)
print(data_array.grad)

输出：

None #<- for the threshold x
tensor([[ 0.1088, -0.0617],  #<- for the data_array
        [ 0.1011,  0.0000],
        [ 0.0000,  0.0000],
        [-0.0000, -0.0000],
        [ 0.2047,  0.0973],
        [-0.0000,  0.2197],
        [-0.0000,  0.0929],
        [ 0.1106,  0.2579],
        [ 0.0743,  0.0880],
        [ 0.0000,  0.1112]])

在 PyTorch 中使用阈值进行训练

Training with threshold in PyTorch

python

machine-learning

deep-learning

pytorch

gradient-descent