Have you found counter-measures for hallucination on this model?
#11
by
jcvasquez
- opened
I have found in the past the output of granite-vision-3.2 is very inconsistent for my use cases and that might be due to its llm part. I think LLM part is the culprit because it has one of the worst hallucination rates in vectara HHEM benchmark hosted in HF (https://huggingface.co/spaces/vectara/leaderboard) up to 15.7% while top models are belo 2% mark and even some older granite like 3.0.8B has only a third of that rate with 6%.