train my own data

[justinner]

Hi,DrSleep,
I want to train model using my own dataset,and I just want to 2 number class,when I changed the parameters in config.py,an error occured as flowing:

size mismatch for module.clf_conv.bias: copying a param with shape torch.Size([40]) from checkpoint, the shape in current model is torch.Size([2]).

in addition,when my label image is  binarized image,is the reason that the error occured?

ValueError: operands could not be broadcast together with remapped shapes [original->remapped]: (2,2) and requested shape (3,2).

hope your help,thanks very much.

[DrSleep]

size mismatch for module.clf_conv.bias: copying a param with shape torch.Size([40]) from checkpoint, the shape in current model is torch.Size([2]).

this indicates that you are trying to load in the checkpoint pre-trained on NYU (with 40 classes).

in addition,when my label image is binarized image,is the reason that the error occured?
ValueError: operands could not be broadcast together with remapped shapes [original->remapped]: (2,2) and requested shape (3,2).

What do you mean by binarised label image? This should not affect anything as long as your masks are of the shape HxW.
Knowing where the ValueError is coming from would be helpful