RuntimeError: Error(s) in loading state_dict for DummyModule:

[ldepn]

你好，我用的python3.7，pytorch 1.4.0 ，在跑demo的时候提示下面这个错误。

RuntimeError: Error(s) in loading state_dict for DummyModule:
size mismatch for module.tl_heats.0.1.weight: copying a param with shape torch.Size([80, 256, 1, 1]) from checkpoint, the shape in current model is torch.Size([21, 256, 1, 1]).
size mismatch for module.tl_heats.0.1.bias: copying a param with shape torch.Size([80]) from checkpoint, the shape in current model is torch.Size([21]).
size mismatch for module.tl_heats.1.1.weight: copying a param with shape torch.Size([80, 256, 1, 1]) from checkpoint, the shape in current model is torch.Size([21, 256, 1, 1]).
size mismatch for module.tl_heats.1.1.bias: copying a param with shape torch.Size([80]) from checkpoint, the shape in current model is torch.Size([21]).
size mismatch for module.tl_heats.2.1.weight: copying a param with shape torch.Size([80, 256, 1, 1]) from checkpoint, the shape in current model is torch.Size([21, 256, 1, 1]).
size mismatch for module.tl_heats.2.1.bias: copying a param with shape torch.Size([80]) from checkpoint, the shape in current model is torch.Size([21]).
size mismatch for module.br_heats.0.1.weight: copying a param with shape torch.Size([80, 256, 1, 1]) from checkpoint, the shape in current model is torch.Size([21, 256, 1, 1]).
size mismatch for module.br_heats.0.1.bias: copying a param with shape torch.Size([80]) from checkpoint, the shape in current model is torch.Size([21]).
size mismatch for module.br_heats.1.1.weight: copying a param with shape torch.Size([80, 256, 1, 1]) from checkpoint, the shape in current model is torch.Size([21, 256, 1, 1]).
size mismatch for module.br_heats.1.1.bias: copying a param with shape torch.Size([80]) from checkpoint, the shape in current model is torch.Size([21]).
size mismatch for module.br_heats.2.1.weight: copying a param with shape torch.Size([80, 256, 1, 1]) from checkpoint, the shape in current model is torch.Size([21, 256, 1, 1]).
size mismatch for module.br_heats.2.1.bias: copying a param with shape torch.Size([80]) from checkpoint, the shape in current model is torch.Size([21]).