Seyed Rohollah Hosseyni
Seyed Rohollah Hosseyni
I have the same issue.
This my question as well.
Hi, Have you found a way to train Swin Transformer on a local machine?
Actually I managed to run the code on my machine. You just have to remove codes that are realted to disributed programming and it works fine.
> 7975 Hi. Did you solve the issue? I also got `nan` for `007975.npy` data.
> @Rruixu I also ran into the same problem. Did you solve end up figuring it out? Hi. Did you solve this issue?
> Hello, we had similar results that if we set `if_categorical=False`, the FID score is better. However, for text-to-motion generation, the diversity of generated motion is also important. When `if_categorical=False`...
> I'm not so sure about your cases on the huge FID score diff for sampling. I think you can try printing the final probability before sampling to check whether...
> The in/out-painting inference is done by teacher-forcing the tokens where we want to keep, and let the VAR transformer only generate the other tokens. For in/out-painting, no class condition...
> It is about 0.9 Thanks