Comparing the functional differences with torch.autograd.enable_grad and torch.autograd.no_grad
torch.autograd.enable_grad
torch.autograd.enable_grad()
For more information, see torch.autograd.enable_grad.
torch.autograd.no_grad
torch.autograd.no_grad()
For more information, see torch.autograd.no_grad.
mindspore.ops.stop_gradient
mindspore.ops.stop_gradient(input)
For more information, see mindspore.ops.stop_gradient.
Differences
PyTorch: Use torch.autograd.enable_grad
to enable gradient calculation, and torch.autograd.no_grad
to disable gradient calculation.
MindSpore: Use stop_gradient to disable calculation of gradient for certain operators.