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

experiment details #1

Open
hopelin99 opened this issue Aug 10, 2023 · 1 comment
Open

experiment details #1

hopelin99 opened this issue Aug 10, 2023 · 1 comment

Comments

@hopelin99
Copy link

Thank you for providing such excellent work! I have a few questions I would like to consult with you.

  1. would like to know if you will release the training scripts.
  2. Have you conducted experiments involving the segmentation of different layers of the LLM?
  3. Have you conducted a comparison between fine-tuning with and without the inclusion of the CAGIT dataset?
  4. I noticed that you used different batch sizes for CAGIT and image-text data. Did you perform a two-stage training process, first using image-text data and then fine-tuning with CAGIT?

I'm looking forward to your answers to these questions. Thank you.

@YYJMJC
Copy link
Collaborator

YYJMJC commented Aug 10, 2023

Thank you for your interest and very insightful questions. We are currently in the process of organizing the training code and haven't completed some systematic exploration experiments (e.g., different injection layers, different proportions of CAGIT data and image-text data...) . Once this phase is finished, we will update the paper and release the code. To train the CLORI module, we now perform one-stage training with mixed data.

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