model-eval-be / requirements.txt
Ahmet Kaan Sever
Now backend writes to results file. Used threading for this.
24a5ac7
raw
history blame contribute delete
306 Bytes
fastapi
uvicorn[standard]
# lm_eval==0.4.3
git+https://github.com/osmangurlek/lm-evaluation-harness.git
git+https://github.com/huggingface/[email protected]
python-jose
python-multipart
deepeval
--extra-index-url https://download.pytorch.org/whl/cu113
huggingface-hub>=0.29.1
torch
sentencepiece