About the result

[Wangld5]

I test the model in dress dataset, and get the result above. The results do not match the paper and I run it following the command you write. Do you know why ?

[numpee]

Hi,
Please refer to the "Notes on Evaluation" part of the README: https://github.com/postBG/CosMo.pytorch#warning-notes-on-evaluation
Also another user pointed out that some papers concatenate all three training subsets, then evaluate on each subset separately. Doing so will actually lead to better performances than what we reported in our paper.

[Wangld5]

Thanks, I will try to uncomment the code presented in fashionIQ.py.