Closed jakubLangr closed 4 years ago
Your agnostic has incorrect shape. It should have 22 channels but seems to have 24. Can you print its shape?
i encounter the same problem as you,how did you fix it?
It has been a while, but AFAIK, it was a case of flattening one of the semantic maps.
Hope it helps!
can you expand on solving this error?
@bhavyashahh ,I solved it by adding a line after https://github.com/sergeywong/cp-vton/blob/fbaf333ba63ddb9c9379e5689241b122f5cf621d/cp_dataset.py#L71
im_parse = im_parse.convert('P')
or maybe you can print the shape of every component compared with the input image proposed by author.
Hope it helps you.
When running the model on my own data and making the change as in #7 (permuting and unpermuting size dimension for
im
andim_h
) I get the following error:This was run after successfully running the model in the first instance. The model currently looks as follows:
The
weight of size [64, 22, 4, 4]
seems predefined by:The input
expected input[4, 24, 256, 192] to have 22 channels
comes from the data size, which for some reason is 2 channels larger?