【报错】RuntimeError: Trying to backward through the graph a second time (or directly access saved tenso

错误代码行定位：

loss.backward()

报错：
RuntimeError: Trying to backward through the graph a second time (or directly access saved tensors after they have already been freed). Saved intermediate values of the graph are freed when you call .backward() or autograd.grad(). Specify retain_graph=True if you need to backward through the graph a second time or if you need to access saved tensors after calling backward.

错误解释：
这个错误表明你在尝试对同一个计算图进行多次反向传播（backward），而 PyTorch 默认情况下只允许对每个计算图进行一次反向传播。在调用 backward() 之后，计算图中的中间张量会被释放，以节省内存。如果你需要对同一个计算图进行多次反向传播，或者在调用 backward() 之后还需要访问这些中间张量，你需要设置 retain_graph=True。

修改：

loss.backward(retain_graph=True)