Machine Learning Research Framework
5.6k 2026-04-18
facebookresearch/mmf
A modular PyTorch-based framework from Facebook AI Research for state-of-the-art vision and language multimodal AI research.
Core Features
Modular architecture for vision and language multimodal research
Reference implementations of state-of-the-art models
Supports distributed training with PyTorch
Acts as a starter codebase for vision and language challenges
Detailed Introduction
MMF (Multimodal Framework) is an open-source, modular framework developed by Facebook AI Research (FAIR) to advance vision and language multimodal AI research. Built on PyTorch, it offers a scalable and fast platform with reference implementations of cutting-edge models. MMF serves as a robust foundation for researchers to bootstrap new projects, participate in challenges, and explore complex interactions between visual and linguistic data, facilitating rapid experimentation and development in the multimodal AI domain.