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.

Sign up for GitHub

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

Jump to bottom

I would like to ask if it’s realistic for training one epoch on a single video to take around 3 minutes. Is this accurate? #58

Open

zzyfight0703 opened this issue Nov 5, 2024 · 1 comment

zzyfight0703 commented Nov 5, 2024

I would like to ask if it’s realistic for training one epoch on a single video to take around 3 minutes. Is this accurate?

Owner

johndpope commented Nov 5, 2024 •

edited

Loading

I’ve been reworking the data loader - this is using overfitting example to one video that does work with latest code fixes in VASA
johndpope/VASA-1-hack#20
(There’s a successful run with that single video)
https://wandb.ai/snoozie/megaportraits/workspace

I limit frames to 100 - so one epoch is just 100 frames with 1 video. I’m sitting on 35,000 videos that I want to train with.

there’s another file in vasa /data/driving.json that that has 4 videos
(Currently investigating this ID leakage)
https://wandb.ai/snoozie/megaportraits/runs/zhos6yxx?nw=nwusersnoozie

johndpope/VASA-1-hack#25

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment