multimodal-training

1 posts with this tag

LoongForge: Scalable Training Framework for Multimodal Transformers

LoongForge: Scalable Training Framework for Multimodal Transformers

LoongForge offers a robust training framework for large-scale transformer models that support language, vision-language, vision-language-action, and diffusion tasks. As part of Baidu's open-source Loong family, it extends Megatron-LM with enhancements for better memory management, component parallel

Administrator 5/6/2026