Move vars to dynamic, add metrics (#1085)

Use pino for logs (#1086)

Kill app on tokenizer load fail (#1044)

Add logs to tokenizer fetch (#1030)

Add langserve endpoint (#1009)

Move default template so it doesn't override tokenizer (#987)

Set ChatML as default chat prompt template (#985)

Add support for cohere endpoints (#976)

Use jinja template for chat formatting (#730) (#744)

Implement Cloudflare Workers AI endpoint (#907) (#972)

Google Vertex API support (#950)

Count system prompt tokens (#850)

Anthropic Endpoint Support (#923)

Add gemma model to prod config (#854)

Models page (#849)

Make lorax openai endpoint working (#755)

[SmallModel] Nit update the comment (#726)

Add embedding models configurable, from both transformers.js and TEI (#646)

Use JSON5 for parsing all config vars (#671)

Add feature to unlist a model (#625)

Standardize HF_ACCESS_TOKEN -> HF_TOKEN (#610)

Add ollama endpoint support (#569)

Add support for tgi multimodal models (#531)

fix: Rename sagemaker endpoint to aws endpoint

Modular backends & support for openAI & AWS endpoints (#541)

Switch chat model back to mistral, use zephyr for small tasks (#515)

Update websearch prompting & summary prompting (#503)

Pass stop sequences properly to TGI (#499)

move some tasks to small model (#479)

Revert "Improve error message when the .env MODELS is not well formatted (#451)"

Improve error message when the .env MODELS is not well formatted (#451)

[Websearch] update (#427)

🐛 Fix - Remove the min length validation on preprompt

Support custom system prompts from the user (#399)

Make all prompt templates configurable (#400)

Add Sagemaker support (#401)

allow different user and assistant end-token (#375)

Add ability to define custom model/dataset URLs (#347)

Disable models (#187)

✨ Add model id to be able to switch models while keeping conversations valid (#181)

make models description dynamic (#169)

🐛 Stop bigcode prompts (#170)

Models selector (#164)

♻️ Simplify setup, prepare for multi-models (#156)