fix: remove Ollama from Dockerfile, use Groq API instead

feat: replace Ollama with Groq API (llama-3.3-70b-versatile)

perf: reduce num_ctx 8192→2048 for faster CPU inference on t3.large-HF

update backend for AWS

Fix: Add zstd dependency for Ollama installation

Add application file

Add application file

Add application file

Add application file

update backend dockerfile - HF

changes for backend - HF

update backend - HF

update backend dockerfile - HF

update dockerfile - HF

update dockerfiles & requirements - HF

update backend for HF

update backend

update

Update app/main.py

update

Fix Hugging Face Space configuration

Initial clean commit with LFS models