Phind new model just beat WizardCoder-Python-34B-V1.0 human eval high score
#13
by
rombodawg
- opened
They got 73.8 vs WizardCoders 73.2. Not a big jump, but now they are competing to be the best open source model out there!
https://huggingface.co/Phind/Phind-CodeLlama-34B-v2
Oh dayum!
Rombodawg, Phind may have stolen WizardLM's model!
This commit happened right after a discussion about the suspicious config.json file. I would link the discussion, but they have since deleted the entire model and reuploaded it. I am guessing because they wanted the discussion and commit history gone.
WizardLM
changed discussion status to
closed