model.eval()
trace_script_module = torch.jit.script(model, (image1, image2))
trace_script_module.save("a.pt")
meet runtimeerror: strides() called on an undefined Tensor
I try to move model and input to cpu, but has no change. And I delete trace_script_module.save("a.pt")
, there is no error, so just the .save()
function error.
But I also can't find where used strides()
.