justlike-prog

I tried with logits, but it is hard to set a threshold there, since it is not normalised (or at least I think this is the issue), so something similar...

Yes, normalisation is applied. Again softmax doesn't work here well, with more non mutual exclusive classes this will be hard to adjust.

Yeah, was thinking about that thanks! I will give it a try. Although online I didn't see examples for calibration in multi label setting.

Ok I will check it out, thanks!

@SimJeg Do you maybe know a guide somewhere about doing such a calibration? Never did anything like that. Thanks!

@mehdidc hmm, this one seems to work wit sklearn models rather than torch for example?

Alright, will do that. Thanks a ton! Would I get a probability for each class independently? I would for example calibrate the scores for class elephant by using the logits...

Is there a direct method to access the weights? Could you share the size in MB of the vision encoder on EC-ConaCLIP-2L so I can see if the size fits...

I have the same problem