This example shows how to use MPT-7B-Chat model, with a SimpleAI server.
It implements chat
methods for both streaming and non-streaming.
First build the image with:
docker build . -t mpt-7b-chat:0.1
Then declare your model in your SimpleAI configuration file models.toml
:
[mpt-7b-chat]
[mpt-7b-chat.metadata]
owned_by = 'MosaicML'
permission = []
description = 'MPT-7B-Chat is a chatbot-like model for dialogue generation. Built by finetuning MPT-7B on the ShareGPT-Vicuna, HC3, Alpaca, Helpful and Harmless, and Evol-Instruct datasets.'
[mpt-7b-chat.network]
type = 'gRPC'
url = 'localhost:50051'
Just start your container with:
docker run -it --rm -p 50051:50051 --gpus all mpt-7b-chat:0.1
And start your SimpleAI instance, for instance with:
simple_ai serve [--host 127.0.0.1] [--port 8080]