workwise-backend-gpu / requirements.txt
VcRlAgent's picture
Generator Refactor HF Inference Client API
a1544bb
raw
history blame contribute delete
625 Bytes
huggingface-hub>=0.26.0
fastapi==0.109.0
uvicorn[standard]==0.27.0
python-dotenv==1.0.0
python-multipart==0.0.6 # if you accept file uploads
# === Data / utilities ===
pandas==2.2.0
numpy==1.26.3
requests==2.31.0
pydantic==2.5.3
tqdm==4.66.4 # progress bars
loguru==0.7.2 # clean logging
# Vector/Embedding stack (keep exactly what you use)
faiss-cpu==1.7.4 # keep if you use FAISS locally
# === ML / AI stack (GPU-compatible) ===
# Using torch for embeddings/models:
torch #-- choose the right wheel for your CUDA (see Dockerfile notes)
transformers
sentence-transformers