license consult
If I need mention the model and dataset origin if I use your model in my videos which will distribute on youtube?
You can mention the LatentSync with the following lines:
Code: https://github.com/bytedance/LatentSync
Model: https://huggingface.co/chunyu-li/LatentSync
Thanks for your reply. Can I only attribute the two train dataset? because they are cc by 4.0 license.
And your pretrain model is Openrail license, and source code is Apache license, which are no need to attribute.
Okay, no problem
After checking the two dataset I noticed that many YouTube videos were used in training the model. I want to avoid any potential copyright issues. Can you tell me if using this model could lead to any copyright disputes?
Many other open source lip-sync models also use Youtube videos for training, so I think it's OK?
For example: MuseTalk, VideoReTalking, DINet