fix: update Groq model list (llama3-70b deprecated -> llama-3.3-70b-versatile)
e228496
DJAKOUA KWANKAM BRAYAN STEVEcommited on
feat: multi-provider (HF Inference + Groq fallback), configurable via env vars
66b89a3
DJAKOUA KWANKAM BRAYAN STEVEcommited on
fix: use router.huggingface.co/hf endpoint (api-inference DNS not resolvable from Space)
ecf2f73
DJAKOUA KWANKAM BRAYAN STEVEcommited on
Switch to free HF Inference API (requests direct), removed Together AI provider. Added requests dep, model fallback, rate limit handling.
aabbe2e
DJAKOUA KWANKAM BRAYAN STEVEcommited on
chore(backend): make Qwen2.5-72B-Instruct default and robust fallback to Llama-3.3-70B-Instruct
9fd7e03
DJAKOUA KWANKAM BRAYAN STEVEcommited on
feat(backend): update default model to Qwen2.5-72B and add Llama-3.3-70B fallback
0ae730f
DJAKOUA KWANKAM BRAYAN STEVEcommited on
style(chat-app): add settings back button, premium dropdown select arrows, browser mockup for preview, update free CPU models list, and refine discussion list cards