Movatterモバイル変換


[0]ホーム

URL:


Skip to content

Navigation Menu

Search code, repositories, users, issues, pull requests...

Provide feedback

We read every piece of feedback, and take your input very seriously.

Saved searches

Use saved searches to filter your results more quickly

Sign up

A wrapper of LLMs that biases its behaviour using prompts and contexts in a transparent manner to the end-users

License

NotificationsYou must be signed in to change notification settings

citiususc/Smarty-GPT

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

69 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

A wrapper of LLMs (chatgpt, gpt4, etc.) that biases its behaviour using prompts

DownloadsPyPi

A wrapper of LLMs that biases its behaviour using prompts and contexts in atransparent manner to the end-users.

Installation

sh install.sh

BinderGoogle ColabGitHub Codespaces

Models

  • text-davinci-003
  • Flan-T5 powered by Google.
  • ChatGPT andGPT4 through paid API.

Contexts / Prompts

We support three type of prompts from the moment:

  • Manual prompts: these prompts are hard-coded and were the first included in this project.
  • Awesome Chat GPT prompts: our system also supports this huge HF dataset in a transparent manner.
  • Custom prompts: any user can add custom prompts through a file.
  • (In progress) Support forawesome-gpt4 prompts.

Authentication

Users should create aconfig.txt file like the following to read Open AI bearer:

[auth]api_key = xxxxxxxxxxxxxxxxxx

Coding examples

fromsmartygptimportSmartyGPT,Modelsif__name__=="__main__":s=SmartyGPT(prompt="DoctorAdvice",config_file="/home/user/config.txt")result=s.wrapper("Can Vitamin D cure COVID-19?")print(result)

Check theColab or test folder for more examples and functionalities

Philosophy

The main purpose of this project isjoining in a single environment all the resources (models, prompts, APIs, etc.) related to LLMs.

Moreover, we also think from anend-user perspective. It is heavily unlikely that a user would introduce a complex context in a query to bias a model response. This library tries to solve this hidding the implementation details to end-users.

In progress

More features/models are about to come! Feel free to make a PR, open an issue or to contact me atmarcosfernandez.pichel@usc.es

Disclaimer

The software is provided "as is" and "with all faults" without warranties of any kind, either express or implied, including, but not limited to, the implied warranties of merchantability, fitness for a particular purpose and non-infringement. No warranty is provided that the software will be free from defects or that operation of the software will be uninterrupted. Your use of the software and any other material or services downloaded or made available to you through the software is at your own discretion and risk, and you are solely responsible for any potential damage resulting from their use.


[8]ページ先頭

©2009-2025 Movatter.jp