Awaker2.5-VL: Stably Scaling MLLMs with Parameter-Efficient Mixture of Experts Paper • 2411.10669 • Published Nov 16, 2024 • 10