Genoss GPT

Genoss GPT is an innovative open-source initiative that offers an alternative to OpenAI's GPT 3.5 and GPT 4 models.

It aims to provide a seamless replacement for OpenAI's ChatGPT API, powered by open-source models like GPT4ALL. 

The project allows users to run language model (LLM) models locally, and it supports various LLMs, both private and public, including models from Anthropic, Llama V2, Vertex, HuggingFace, and others.

At its core, Genoss GPT is built on top of open-source models like GPT4ALL and can be easily installed on Python 3.11 or higher. The application utilizes the Uvicorn ASGI server to run the Genoss API on a user's local machine. It offers support for running text completion and embeddings using the GPT4ALL model.

The vision behind Genoss is to expand its offerings by adding support for more models in the future. While currently focused on GPT4ALL, the project plans to accommodate various LLMs, including text-to-text, text-to-image, text-to-audio, audio-to-text, and more.

The Genoss API serves as a one-line replacement for the OpenAI ChatGPT API, making it easy for users to transition from OpenAI's models to Genoss GPT seamlessly. The API documentation can be found at /docs or /redoc.

Genoss GPT was conceived by Stan Girard, and it was brought to life with contributions from developers @mattzcarey, @MaximeThoonsen, @Wirg, and @StanGirard. The project is licensed under the Apache2 License and welcomes contributions from the community.

Overall, Genoss GPT offers an exciting alternative for developers and researchers to utilize powerful language models while maintaining the flexibility of open-source solutions.