Question

我目前正在尝试用pytorch训练神经网络，在这里我尝试匹配输入导数上的输入。我要这样做是因为这确保了保守的矢量场。（完成神经网络训练以进行分子动力学力匹配）这意味着：

input = torch.rand((n,3),requires_grad=True)
output = torch.rand((n,3),requires_grad=True)
prediction = model(input) # size of the prediction[1]
input_grad = torch.autograd.grad(outputs=prediction, inputs=input, retain_graph=True, create_graph=True)
loss = loss_fn(output,input_grad)
...

问题是，如果我尝试更新神经网络的参数，则所有参数的梯度均为0。我不知道如何建立模型正在正确训练的图形。在Jaxmd中，可以像[Jax Glass Training] [1]所示训练这样的模型。我也尝试过

input_grad = torch.autograd.grad(outputs=prediction, inputs=input, create_graph=True)

但是这会产生相似的结果并且没有任何意义。 [1]：https://colab.research.google.com/github/google/jax-md/blob/master/notebooks/neural_networks.ipynb#scrollTo=WNs8v2745Mc3

编辑：

再现pytorch版本1.6.0的更新代码示例

import torch 

class Feedforward(torch.nn.Module):
        def __init__(self, input_size, hidden_size):
            super(Feedforward, self).__init__()
            self.input_size = input_size
            self.hidden_size  = hidden_size
            self.fc1 = torch.nn.Linear(self.input_size, self.hidden_size)
            self.relu = torch.nn.ReLU()
            self.fc2 = torch.nn.Linear(self.hidden_size, 1)
     
        
        def forward(self, x):
            hidden = self.fc1(x)
            relu = self.relu(hidden)
            output = self.fc2(relu)
            output = output.sum()
            output =torch.autograd.grad(outputs=output, inputs=x, retain_graph=True, create_graph=True)
            return output[0]

test_input = torch.rand((10,3),requires_grad=True)
test_output = torch.rand((10,3))


model = Feedforward(3,10)
optim = torch.optim.Adam(model.parameters())
optim.zero_grad()
loss_fn = torch.nn.L1Loss()
model.train()
out = model(test_input)
loss = loss_fn(out, test_output)
loss.backward()
optim.step() # if you break here and investigate the gradients
             # of the FFNN, the gradients will be 0

用pytorch训练输入梯度的神经网络

0 个答案: