Search papers, labs, and topics across Lattice.
Peking University, Tencent Hunyuan
3
0
4
GEAR accelerates image synthesis convergence by up to 10x while enhancing feature coherence, challenging the traditional decoupling of tokenizers and generators.
Forget scalar rewards: GenEvolve distills structured visual experiences from successful and failed generation trajectories, enabling token-level supervision for self-improving image generation agents.
OpenSearch-VL offers a fully transparent recipe for training state-of-the-art multimodal search agents, finally democratizing access to a capability previously locked behind closed doors.
