Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Is there a simplified training method? #81

Open
zhjygit opened this issue Jan 15, 2025 · 1 comment
Open

Is there a simplified training method? #81

zhjygit opened this issue Jan 15, 2025 · 1 comment

Comments

@zhjygit
Copy link

zhjygit commented Jan 15, 2025

Is there a simplified training method?
My GPU 4060ti has 16 GB of video memory and more than 100 GB of datasets, can I save time during training by simplifying the steps or reducing the number of datasets I train? In the training phase, if it is for a certain language, you only need to train the u-set, right?

@zhjygit
Copy link
Author

zhjygit commented Jan 15, 2025

Downloaded Chinese datasets, which are video, audio, and train_csv val.csv, and there are a large number of mp4 format videos in the video and audio folders, and the length of each video is less than 3 seconds. In this case, it seems that it is not possible to directly use ./data_processing_pipeline.sh for data processing, and all videos will be automatically deleted, which may not meet some of the reading requirements of AVREARDER, and the video length does not meet the requirements. If it is training in Chinese, can you provide some dataset links and training steps?

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

1 participant