logikon-ai / cot-eval

A framework for evaluating the effectiveness of chain-of-thought reasoning in language models.
https://huggingface.co/spaces/logikon/open_cot_leaderboard
MIT License
12 stars 2 forks source link

Evaluate: upstage/SOLAR-10.7B-v1.0 #28

Closed ggbetz closed 7 months ago

ggbetz commented 8 months ago

Check upon issue creation:

Parameters:

NEXT_MODEL_PATH=upstage/SOLAR-10.7B-v1.0
NEXT_MODEL_REVISION=main
NEXT_MODEL_PRECISION=float16
MAX_LENGTH=2048 
GPU_MEMORY_UTILIZATION=0.8
VLLM_SWAP_SPACE=4

ToDos: