/mtllm

The MTLLM Python library provides convenient access to a large number of easy to use and customizable APIs to be used in Jaseci's Jaclang by llm feature. The Library provides automatic output fixing, output type validation, different prompting techniques, and more.

Primary LanguagePythonMIT LicenseMIT

MTLLM API library

PyPI version tests

The MTLLM Python library provides convenient access to a large number of easy to use and customizable APIs to be used in Jaseci's Jaclang by llm feature. The Library provides automatic output fixing, output type validation, different prompting techniques, and more.

Documentation

The documentation on how to use this library with Jaseci's Jaclang can be found here.

Installation

Important

Though this is can be used with python projects, it is primarily intended to be used with Jaseci's Jaclang.

# install from PyPI
pip install mtllm

Usage

Refer the Documentation for detailed usage instructions.

Using Different LLMs

import:py from mtllm.llms, OpenAI;

glob llm = OpenAI();

can "Translate English to French"
translate(word: "English Word": str) -> "French Word": str by llm();

Based on your LLM of choice, make sure to set the API Key in the environment variable. For example, for OpenAI:

export OPENAI_API_KEY="your-api-key"

Using Tools

import:py from mtllm.llms, OpenAI;
import:py from mtllm.tools, wikipedia;

glob llm = OpenAI();

can "Answer History Questions"
history_qa(question: "History Question": str) -> "Detailed Answer": str by llm(tools=[wikipedia]);

Contributing

Please read CONTRIBUTING.md for details on our code of conduct, and the process for submitting pull requests to us.