JialeTao / MTIA

[ECCV 2022] Motion Transformer for Unsupervised Image Animation

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

a issue about demo

Explore008 opened this issue · comments

You have done a great job! But when I tried to run the demo file, I encountered an error that was displayed as' load '_ A parameter is missing from checkpoint. Could you please tell me how to solve it? Thank you!
1692270938623

Hi @Explore008 , this may be caused by the lib version of "yaml". using the command "yaml.load(file, Loader=yaml.FullLoader)" may helps.

Hi @Explore008 , this may be caused by the lib version of "yaml". using the command "yaml.load(file, Loader=yaml.FullLoader)" may helps.
Thank you for your reply! Your help is very effective, but when I drive the image character by loading the pre-trained taichi model, it produces great distortion and confusion, unlike a person. I wonder what caused it. Perhaps it is the clarity of the video and the pixel size of the image? I would appreciate it if you could provide a suitable video.

1.mp4
3.mp4

The image and video size may not influence. This may be caused by the inability of the motion estimator for cross-domain motion estimation. To verify this and for better understanding, you could visualize the motion masks and keypoints of the driving video and source image.

图像和视频大小可能不会影响。这可能是由于运动估计器无法进行跨域运动估计造成的。为了验证这一点并更好地理解,您可以可视化驾驶视频和源图像的运动遮罩和关键点。

Sorry,I can not find this problem.I want to experiment with taichi data set to see your excellent scientific research results. However, the download link of this file is wrong and the data is too large to load easily. Can you give me a copy that you have handled? I found your github email, but its domain name is not available. If you can, you can share it with my academic mailbox, 202233374@stumail.nwu.edu.cn. Thank you very much!

You can send an email to me (jialetao.std@gmail.com) and comment on your academic purpose, then I can share a copy of the dataset.

You can send an email to me (jialetao.std@gmail.com) and comment on your academic purpose, then I can share a copy of the dataset.

I had sent a message to your email, but this message is refused by your service.

Done it.

You can send an email to me (jialetao.std@gmail.com) and comment on your academic purpose, then I can share a copy of the dataset.

I had sent a message to your email, but this message is refused by your service.