Ervin Áron Tasnádi
Ervin Áron Tasnádi
Thanks, then I will remove the assert if you think it is not needed anymore. I thouht that the code is not functioning properly on my system as there is...
I rebuilt in Release and it does not crash anymore (not in the cases mentioned at least).
It is the cosine similarity in the paper that is usually computed using dot product, therefore bmm may compute the cos-similarity of each possible vector pairs ([cosine] similarity matrix) coming...
> Hello, I have the same quetion, in fact, I trained the model on the COCO dataset, the loss seems to quickly goes to -2 and stagtnates there The same...
Add `#include ` for the corresponding sources. Or switch to `gcc-9`.
Dear All, If anyone is considering to implement this in torchvision, I have a working 3D RoiAlign kernel implemented in Tensnorflow that could be directly ported back into PyTorch. You...
Might worth considering https://github.com/TimothyZero/MedVision/tree/main for the torch version also.
> > Corrected test drive in sd.cpp for VAE decoding: > > ### 768x1024 sd1 fp16 vae: > > method time memory > > CUDA imcol+mul ~1.68s 4992.19 MB >...
> Now #16088 is a better implementation. Actually, it is a somewhat better implementation you just forget to enable it in your code so the previous inefficient implementation was used,...