size mismatch error during inference using the provided guacamole checkpoint
chetanrrk opened this issue · comments
I am trying to run inference using the provided checkpoint file for guacamole. As pointed in the config, I change the test_only to the absolute path of the check point file. Also, as suggested, I am reverting to the "682e59019dd33073b1f0f4d3aaba7de6a308602e" state. I get the following trace:
"""
raise RuntimeError('Error(s) in loading state_dict for {}:\n\t{}'.format(
RuntimeError: Error(s) in loading state_dict for DiscreteDenoisingDiffusion:
size mismatch for model.mlp_out_y.2.weight: copying a param with shape torch.Size([1, 256]) from checkpoint, the shape in current model is torch.Size([0, 256]).
size mismatch for model.mlp_out_y.2.bias: copying a param with shape torch.Size([1]) from checkpoint, the shape in current model is torch.Size([0]).
"""
Looks like a bug. Any suggestion? Thanks!
Hello,
could you try to change the output dimension of y to 1 in the code ? The output value will never be used, but it's not really a problem.