Difference with h2oai/h2ogpt-oig-oasst1-falcon-40b
#2
by
tomheno
- opened
Hey !
I was wondering what was the difference in dataset between this model and the h2oai/h2ogpt-oig-oasst1-falcon-40b ?
What is the expected difference in output between the two ?
Thanks in advance :)
This model was trained on a different dataset. For MMLU evaluation metrics, see here: https://github.com/h2oai/h2ogpt/issues/251#issuecomment-1585183601
Related question:
Also, what was the difference between those two and https://huggingface.co/h2oai/h2ogpt-gm-oasst1-en-2048-falcon-40b-v1 ?
Our current best model is: https://huggingface.co/h2oai/h2ogpt-gm-oasst1-en-2048-falcon-40b-v2 (gm stands for Kaggle Grandmasters)
arnocandel
changed discussion status to
closed