Local LLM Server with NPU Acceleration

Function-calling API for LLM from multiple providers

Simple LLM Rest API using Rust, Warp and Candle. Dedicated for quantized version of either phi-2 ( default) , Mistral, or Llama. Work using CPU or CUDA

A flexible FastAPI-based framework for handling AI tasks using Large Language Models (LLMs). Supports multiple providers, extensible tasks and routers, Redis caching, and OpenAI integration. Easily scalable for various LLM-based applications.