MHA2MLA-VLM: Enabling DeepSeek's Economical Multi-Head Latent Attention Across Vision-Language Models

Xiaoran Fan, Zhichao Sun, Tao Ji, Lixing Shen, Tao Gui

BibTeX
@inproceedings{fan2026mha,
  author = {Xiaoran Fan and Zhichao Sun and Tao Ji and Lixing Shen and Tao Gui},
  title = {{MHA2MLA-VLM: Enabling DeepSeek's Economical Multi-Head Latent Attention Across Vision-Language Models}},
  booktitle = {AAAI Conference on Artificial Intelligence},
  year = {2026},
  doi = {10.1609/aaai.v40i36.40319},
}
Figure