leeroo-client 0.0.1

A client library for Leeroo Workflow Management

Leeroo Client

Leeroo Dager offers a comprehensive solution for developing custom AI models. By simply providing a task description and seed data, Leeroo automates the entire workflow from data generation to model training and evaluation. No AI expertise is required—Leeroo ensures the delivery of the best customized model tailored to your specifications. Model deployment is streamlined and can be achieved with a single command.

Installation

Install using pip

pip install leeroo-client

Install from Source

git clone https://github.com/Leeroo-AI/leeroo-client
cd leeroo-client
pip install -e .

Features

Synthetic Data Generation
Our advanced data generation pipeline employs multi-agent systems to create diverse training data, ensuring enhanced generalization and reliability.

Training
Our training pipeline is designed to execute optimal experiments tailored to your use case on any cloud provider. By integrating state-of-the-art techniques such as SFT, DPO, RLHF or building compound models, we consistently deliver high-performing custom models.

Evaluation
Utilize our Multi-Model LLM System to create high-quality LLM as judge Evaluation systems.

Deployment
Automatically deploy customized AI models without the need for AI Infra expertise. Uses VLLM / Fastchat for deployment.