Weizhi-Zhong / IP_LAP

CVPR2023 talking face implementation for Identity-Preserving Talking Face Generation With Landmark and Appearance Priors

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

the lip shakes heavily after training

MYTHOFLEGEND opened this issue · comments

Hello, I trained the landmarks model by the videos I collected from bilibili, running_L1_loss is 0.0066 and running_velocity_loss is 0.0048. but in the inference result, the lip shakes heavily and mismatch with the upper face

4result_N_25_Nl_15.mp4

could you give me some suggestions to fix that? Thanks very much!

同样存在这个问题,不过我安装的一些库可能跟作者在requirements列的不太一样,不知道是不是这个原因。请问你是按照作者给的requirements安装的环境吗?

不是,我都是尽量装的最新,pytorch用的2.0.1,直接用作者的模型推理就没事,自己训练的就会出现这个问题

我用预训练模型都会抖

@MYTHOFLEGEND hello bro!! how to get separate video or separate video for show_sketch and separate video for full

do you have any ideas to improve resolution?

commented

@MYTHOFLEGEND Hi! If possiable, Can you tell me what you used to remove the background in your video.I'm very interested in this.Thanks in advance.

上下对不上可能是因为推理和训练时 crop 人脸时设置的 pad 不一致导致的,这个代码里处理数据时是 5,推理时是 25