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

[Question] About the training data llava_v1_5_mix665k_clean_ok.json #4

Open
jungle-gym-ac opened this issue Aug 10, 2024 · 1 comment

Comments

@jungle-gym-ac
Copy link

Question

Hi, great work!
I noticed that the training data path in the training script you provided is llava_v1_5_mix665k_clean_ok.json, which is not the original llava_v1_5_mix665k.json . Did you do any data cleaning or post-processing to the original json provided by llava?
Thank you!

@senfu
Copy link
Contributor

senfu commented Aug 24, 2024

We are still organizing our training data, and will release them soon. Basically we just delete those data that has invalid image (i.e. an empty image by some reasons) and rewrite the text only data so that they can fit into our training pipeline. The data we are using is mostly the same as the original llava_v1_5_mix665k.json.

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

2 participants