search

Weizhi-Zhong / IP_LAP

CVPR2023 talking face implementation for Identity-Preserving Talking Face Generation With Landmark and Appearance Priors
Apache License 2.0
665 stars 73 forks source link

train, inference code and pretrained weight #3

Closed KangweiiLiu closed 1 year ago

KangweiiLiu commented 1 year ago

Thank you for your hard work on this project! I'm excited to see the code and would love to know when it will be made available to the public. Do you have an estimated timeline for when this might happen?

Weizhi-Zhong commented 1 year ago

Hi~, thanks for attention. As depicted in README.md, we plan to complete the project(including train, inference code and pretrained weight) before May 15th. We are trying our best to reorganize the code and make it more readable these day.

KangweiiLiu commented 1 year ago

errors in inference: RuntimeError: Error(s) in loading state_dict for Landmark_generator: Missing key(s) in state_dict: "lip_landmark_map.weight", "lip_landmark_map.bias", "jaw_landmark_map.weight", "jaw_landmark_map.bias". Unexpected key(s) in state_dict: "mouse_keypoint_map.weight", "mouse_keypoint_map.bias", "jaw_keypoint_map.weight", "jaw_keypoint_map.bias".

wrong result issue

Weizhi-Zhong commented 1 year ago

errors in inference: RuntimeError: Error(s) in loading state_dict for Landmark_generator: Missing key(s) in state_dict: "lip_landmark_map.weight", "lip_landmark_map.bias", "jaw_landmark_map.weight", "jaw_landmark_map.bias". Unexpected key(s) in state_dict: "mouse_keypoint_map.weight", "mouse_keypoint_map.bias", "jaw_keypoint_map.weight", "jaw_keypoint_map.bias".

wrong result issue

sorry, the reason may be that we rename some variable in the latest code for readability, but our pre-trained models are trained using old code. we will fix this as soon as possible