DianeBouchacourt

Hi, Yes but this is not the only change that has to be done. The implementation of the accumulation of evidence function, and some bugs (e.g. calling .data on Variable...

Has this issue been solved ? Training on dSprites, I also get negative tc loss

After a bit more digging, I think the problem comes from the fact that indexes are selected but the view() operation https://github.com/YannDubs/disentangling-vae/blob/7b8285baa19d591cf34c652049884aca5d8acbca/disvae/evaluate.py#L272 is reordering in a wrong manner, permute is...

Let me check my code again, and I will ! I also need to do a PR just for these changes (I edited quite some stuff in the fork I...

I have multiple images of the same latents (other factors are unknown, and can differ between images but I don't have ground-truth for these, e.g. multiple images of the same...

If you are still interested, the is a nice response on stackoverflow : http://stats.stackexchange.com/questions/250838/integral-identity-of-lemma-contained-in-infogan-paper