OpenAssistant

Quick Experience

openassistant app

darkassistant is the new version(the task will be completed by agents):
darkassistant app

Features

Let's ask the assistant first🤔:

So this project encompasses the following features:

1. Rapid Conversion of LLM to Agent🤖

Effortlessly transform your Language Model (LLM) into an Agent.

2. LLM Proficiency Testing Tool🛠️

Explore and evaluate the capabilities of your Language Model through the integrated testing tool.

3. Open Assistant WebUI

Experience the convenience of the Open Assistant Web User Interface (WebUI).

Watch the demo(Vicuna v1.5)

• Youtube
• BiliBili

Everyone have their own AI assistant

Support models

models on huggingface:

• vicuna
• airoboros
• koala
• alpaca
• chatglm
• chatglm2
• dolly_v2
• oasst_pythia
• oasst_llama
• tulu
• stablelm
• baize
• rwkv
• openbuddy
• phoenix
• claude
• mpt-7b-chat
• mpt-30b-chat
• mpt-30b-instruct
• bard
• billa
• redpajama-incite
• h2ogpt
• Robin
• snoozy
• manticore
• falcon
• polyglot_changgpt
• tigerbot
• xgen
• internlm-chat
• starchat
• baichuan-chat
• llama-2
• cutegpt
• open-orca
• qwen-7b-chat
• aquila-chat
• ...

How to use

1. Installation

git clone https://github.com/Qiyuan-Ge/OpenAssistant.git
cd OpenAssistant
pip install -r requirements.txt

2. Starting the server

First, launch the controller:

python3 -m fastchat.serve.controller

Then, launch the model worker(s):

python3 -m fastchat.serve.multi_model_worker \
    --model-path model_math \
    --model-names "gpt-3.5-turbo,text-davinci-003" \
    --model-path embedding_model_math \
    --model-names "text-embedding-ada-002"

Finally, launch the RESTful API server：

python3 -m fastchat.serve.openai_api_server --host 0.0.0.0 --port 6006

You should see terminal output like：

INFO:     Started server process [1301]
INFO:     Waiting for application startup.
INFO:     Application startup complete.
INFO:     Uvicorn running on http://0.0.0.0:6006 (Press CTRL+C to quit)

See more details in https://github.com/lm-sys/FastChat/blob/main/docs/openai_api.md

3. Starting the web UI

streamlit run main.py

Then replace the API Base with your api base (in this case is http://0.0.0.0:6006/v1)