aryan-01001/knowledgebase

fix: remove HF_HUB_OFFLINE from Dockerfile so build step can download the model

fix: moved Dockerfile to root and updated paths for Hugging Face Spaces

fix: add app_port to HF configuration

fix: rename readme.md to README.md so Hugging Face can detect the config block

hf config

feat: setup Dockerfile for Hugging Face Spaces deployment

fix: change fastembed cache_dir to models to prevent shadowing by persistent data mount

fix: set HF_HUB_OFFLINE=1 to force offline model loading at runtime

update

fix: cap fastembed and FAISS thread count to 1 to stay within 512MB RAM

fix: pre-cache fastembed model at build time to prevent request timeout & container restart

updated readme

fix: update fastembed to 0.8.0 for Python 3.14 compatibility on Render

fix: use fastembed (ONNX) instead of HF API - works within Render sandbox and 512MB RAM

fix: replace local PyTorch models with HF Inference API to resolve OOM on Render

updated reqs

optimized memory

optimized memory

updated ui

updated workflow and ui

Frontend

RAG implement

Initial