PrototypeCommon
DAGroup-PKU/MHLA
[ICLR 2026🔥] MHLA: Restoring Expressivity of Linear Attention via Token-Level Multi-Head
- Owner
- DAGroup-PKU (organization)
- Stars
- Stars: 160
- Forks
- Forks: 6
- Language
- Python
- License
- MIT
- Created
- 2025
- Topics
- ditimage-classificationimage-generationimagenetlanguage-modelinglinear-attentionsanavideo-generation
- Tier
- Prototype (100 to 249 ★) What this tier means
More repos by DAGroup-PKU
All 7 repos by DAGroup-PKU- DAGroup-PKU/HumanNetStars: 296PythonAlpha
- DAGroup-PKU/ResearchBoost-AgentStars: 210PythonPrototype
- DAGroup-PKU/PyRUA-LeanStars: 173PythonPrototype
- DAGroup-PKU/PhysisForcingStars: 136PythonPrototype
- DAGroup-PKU/ViGARStars: 113PythonPrototype
- DAGroup-PKU/ReVidgenStars: 103Jupyter NotebookPrototype
Nearby in Python
All Python repos- Video-Rebirth/hyperflowStars: 160PythonPrototype
- allebee/jevk5Stars: 160PythonPrototype
- zxcHolmes/minimax-h3-studioStars: 160PythonPrototype
- sethtrain/django-gitStars: 160PythonPrototype
- knipknap/GelatinStars: 160PythonPrototype
- niemeyer/hsandboxStars: 160PythonPrototype