requirements.txt

fastapi>=0.111
uvicorn[standard]>=0.30
gradio>=4.38
requests>=2.31

# vLLM & friends
vllm==0.5.5
transformers>=4.44
accelerate>=0.30

# Torch for CUDA 12.1 base image
--extra-index-url https://download.pytorch.org/whl/cu121
torch==2.4.0

# Pin these to avoid the Numba cache error
outlines==0.0.46
numba==0.59.1