The first multimodal generative understanding unified model Lumina-DiMOO based on Ascend MindSpeed MM training is officially released.

At the World Artificial Intelligence Conference in 2025, Shanghai Artificial Intelligence Laboratory, in collaboration with Huawei Ascend, launched a new generation multi-modal generation and understanding unified model Lumina-DiMOO, supporting various tasks such as text to image/video generation, image editing, image translation, and image restoration. This model is the first multi-modal generation and understanding unified model that achieves efficient training throughout the entire process based on the Ascend AI basic software and hardware platform and MindSpeed MM multi-modal large model suite.