Replies: 2 comments 1 reply
-
Thanks for your attention, we move multi-modality model evaluation to this repo: |
Beta Was this translation helpful? Give feedback.
0 replies
-
@bittersweet1999, thank you for a prompt response. I am looking for a way to evaluate LLaVA on text-only benchmarks, not on multi-modal evaluations. |
Beta Was this translation helpful? Give feedback.
1 reply
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
In instruction-tuning, LLaVA trains the LLM as well. I want to evaluate the LLM after the instruction tuning for text-only tasks such as MMLU. How to do this? From what I understand, LLaVA saves the projection layer together with the LLM, which is failing the evaluation attempts. Any workarounds?
Beta Was this translation helpful? Give feedback.
All reactions