# Multimodal Transformation

4M is a 'any-to-any' foundational model training framework that achieves multimodal expansion through tokenization and masking techniques

4M is a framework for training 'any-to-any' multimodal foundation models, extending to various different modalities through tokenization and masking techniques.