error in main_model_params.

[whubaichuan]

hi, nice work.

The optimizer for the main model seems to be wrong. The code in this file:

if all(p is not x for x in model.classification_loss.parameters()) but model.classification_loss.parameters() is empty. Thus, I guess the code should be changed to: if all(p is not x for x in model.linear_classifier.parameters()) ?

[saum7800]

Thanks for pointing that out, @whubaichuan ! What you are saying seems to be right, since classification_loss is a loss fn and not a model, so parameters() would be empty for it. Were you able to test whether the code works with your change? Would be happy to merge your PR making this change if it works. Thanks!

[whubaichuan]

@saum7800 Hi, thanks for your quick reply. I already pulled one request, to solve the empty problem of classification_loss.