This is your work, valued
DualSpeed. Fast-Slow Efficient Training for Multimodal Large Language Models via Visual Token Pruning
