StephanPan

> The lip synthesis effect of auido2exp_00300-model.pth given by the author is not satisfied, can I use the wav2lip model trained by myself, can you tell me what I should...

https://github.com/sicxu/Deep3DFaceRecon_pytorch/blob/b9e6fa89ea1566bb9ec27922f6f5d87c34b4b261/models/bfm.py#L294 face_texture = torch.ones(1,35709,3).float().cuda()*0.7 you could set the face_texture to a pure color here

thanks for your answer! and i wonder whether it is suitable for light-weighted model such as yolov4tiny, and i am in experiment, the result is not well, i simply changed...

i simply change the backbone according to the yolov4tiny, and the anchor of 512 is deleted since the limited img size. btw, the activation is replaced by leakyrelu. other setting...

Thx for your reply! Another question is that wheather the multi-scale training and swa are included?

thx a lot! i find that when i change the test img size from 608 to 320, the performance drops a lot. map drops from 43.2 to 34.5. The performance...

Normally, GPU usage is around 2400 M, but occasionally, it may exceed 10 G.

> Does this UserWarning appear? You can turn it off below > > ``` > UserWarning: operator() profile_node %385 : int[] = prim::profile_ivalue(%383) does not have profile information (Triggered internally...

当推理帧和原始帧差异较大，结果会有双下巴，是需要训一下torso吗？还是贴脸的逻辑不太对？ https://github.com/ZiqiaoPeng/SyncTalk/assets/37864281/540879eb-f86a-436f-8b33-fa7bfb3c22bd

有什么tricks可以提高清晰度吗，感觉预测的图像清晰度相较训练集有所下降？