/SkyChat-CN-Chatbot-GPT3

SkyChat is a Chatbot project based on Chinese GPT3 API. Like chatGPT, it can do human-machine chat, question and answer, and can also complete tasks such as Chinese-English or English-Chinese translation, content continuation, couplets, and Chinese ancient poems writing. / SkyChat是一款基于中文GPT-3 API做的聊天机器人项目。它可以像chatGPT一样，实现人机聊天、问答、中英文互译、对对联、写古诗等任务。

Primary LanguageC#MIT LicenseMIT

SkyChat

简体中文

English

Brief introduction

SkyChat is a chat robot project developed based on the Chinese GPT-3 API. In addition to basic chatting, dialogue, and asking me to answer, it can also support Chinese-English translation, content continuation, couplets, writing ancient poems, generating recipes , third-person reporting, creating interview questions and many other functions.

image

More details at Singularity-AI-OpenAPI

Here are some show cases:

Show cases

For experience and trial, please visit Singularity-AI-trail

Chat

image

Question and answer

image

Generate recipes：

input—— image

output—— image

Couplet

image

How to use the Demo：

Set up models and python environment

Windows

  1. Download Anaconda , and select add to path
    图片名称
  2. Download and unzip the models and training data semantic_score_clean and user_profile_clean
  3. Open an Anaconda prompt and type in the commands
    conda create -n semantic create new env
    conda info --envs check environments
    activate semantic activate it
    cd [YOUR PATH]\semantic_score_clean\semantic_score_clean navigate to the extracted 'semantic_score_clean' folder
    python -m pip uninstall numpy Delete the default numpy
    python -m pip install -r requirements.txt Install Requirements
    python semantic_score_api.py Bring semantic service online
  4. Download Cuda according to your system and install
  5. Open a new anaconda prompt and type in the following conda create -n userprofilecreate new env
    conda info --envscheck environments
    activate userprofileactivate it
    cd [YOUR PATH]\user_profile_clean\user_profile_clean navigate to the extracted 'user_profile_clean' folder
    python -m pip uninstall numpy Delete the default numpy
    python -m pip install -r requirements.txt Install Requirements
    python -m pip install torch==1.11.0+cu115 -f https://download.pytorch.org/whl/torch_stable.html Download torch in cuda version
    python server_v3.pyBring userprofile service online

Unity Demo

  1. Download the newest release
  2. Register In OpenAPI website or inside the Demo to get 'api-key' and 'api-secret'
  3. Click Setting, inside the setting panel, type in key and secret accordingly, customize the names in your need, click save and exit
  4. Make sure 'semantic_score_clean' and 'user_profile_clean' has been set up locally, if you want to set it up remotely, replace the url to your remote host
  5. Time to chat!

Python Demo

  1. Register In OpenAPI website to get 'api-key' and 'api-secret'
  2. Open \[YOUR PATH]\OpenAPIDemo\PythonDemo\main.py fill in key and secret accordingly 图片名称
  3. Open Anaconda and type in the following conda create -n talk create new env
    activate talk activate it
    cd \[YOUR PATH]\OpenAPIDemo\PythonDemo navigate to your Demo path
    python -m pip install requests Install requirements
    python main.py Begin to chat（Make sure 'Semantic' and 'UserProfile' services has been set up）
* Minimum requirement Nvidia Gfx 1060