Xingkai Yu

multimodal understanding (2) mixture of expert (2) vision-language model (2) image generation (2) autoregressive model (2) multimodal learning (2) rectified flow (1) expert specialization (1) knowledge sharing (1) multimodal generation (1) large language model (1) visual encoding (1) sparsity allocation (1) unified multimodal model (1) conditional memory (1) autoregressive framework (1) transformer architecture (1) deep sparse embedding (1) model architecture (1) scaling law (1)