search

NVlabs / imaginaire

NVIDIA's Deep Imagination Team's PyTorch Library
Other
4.01k stars 450 forks source link

Pretrained Few-shot vid2vid for Street model ? #82

Closed DanAmador closed 3 years ago

DanAmador commented 3 years ago

Good day,

the fsvid2vid repo shows a gif with 3 trained models showcasing this architecture. So far only a single pretrained model has been released .

But the one that I currently need is the street one. Is there anywhere I can find more information about the Street model (or the weights themselves)? Or information on how to train it ?

git-hamza commented 3 years ago

@DanAmador Hey I can not find the pose model, can you please share a link where is the pre-trained model of ew-shot vid2vid pose.

jashbari93 commented 3 years ago

Yeah @DanAmador can you please share the pretrained pose model

DanAmador commented 3 years ago

I accidentally said that the pose one was the released one, so far the only fsvid2vid model that has been released is the faceForensics one. I do not possess the YoutubeDancing one and I believe they won't release it thanks to copyright.