fastapi==0.110.3
uvicorn[standard]==0.30.6
httpx==0.27.2
pydantic==2.9.2
pynvml==12.0.0
psutil==6.1.0
python-dotenv==1.0.1
# vLLM nightly instalowany w Dockerfile.cuda (nie tu — heavy CUDA wheels):
# pip install --pre vllm --extra-index-url https://wheels.vllm.ai/nightly
