add support for Mistral using TGI / vllm / candle

[pabl-o-ce]

Hi guys love your project

I was wondering if you can add support to mistral via:

• TGI
• vllm
• candle

for use it as endpoints also they have active support to new llm Architectures as Mistral

[williamhogman]

Hey sounds like a very good idea :)

If anyone wants to add this it would be a most welcome contribution.

[andychenbruce]

Llama+Mistral+Zephyr and GPU acceleration in only ~450 lines using candle.
https://github.com/huggingface/candle/blob/main/candle-examples/examples/quantized/main.rs

If Mistral support is added with candle it could be fairly trivial to also support Llama and Zephyr.

[01PrathamS]

I have some experience with Rust, although my familiarity with LLMS is somewhat limited. can take on this challenge, as it would mark my initial contribution to the LLM-chain.

[williamhogman]

Sounds like a great idea :)