The esteemed Doubao Large Model Team is proud to unveil the technical report on Wenshengtu, offering an unprecedented glimpse into the Seedream 2.0 image generation model. This report meticulously outlines the intricate processes of data construction, the pre-training framework, and the post-training RLHF (Reinforcement Learning from Human Feedback), providing a comprehensive understanding of the technology behind Wenshengtu.
