AI Fine-tuning Tool
2.7k 2026-04-30
roboflow/maestro
A streamlined tool to accelerate the fine-tuning of popular multimodal models like Florence-2, PaliGemma 2, and Qwen2.5-VL.
Core Features
Accelerated fine-tuning for multimodal models
Supports Florence-2, PaliGemma 2, and Qwen2.5-VL
Encapsulates best practices for configuration, data loading, and training
Offers LoRA, QLoRA, and graph freezing for hardware efficiency
Provides both CLI and Python API for flexible usage
Quick Start
pip install "maestro[paligemma_2]"Detailed Introduction
Maestro is an open-source tool designed to simplify and accelerate the fine-tuning process for state-of-the-art multimodal models. It abstracts away complexities related to configuration, data preparation, and training loop setup, allowing developers and researchers to quickly adapt models like Florence-2, PaliGemma 2, and Qwen2.5-VL to specific tasks. By integrating best practices and offering efficient optimization strategies like LoRA and QLoRA, Maestro significantly reduces the effort and hardware requirements for achieving high-quality custom AI models.