MHLA: Restoring Expressivity of Linear Attention via Token-Level Multi-Head
Paper
•
2601.07832
•
Published
•
36
yfdeng
yfdeng10
·
AI & ML interests
None yet
Recent Activity
authored
a paper
about 3 hours ago
CINEMA: Coherent Multi-Subject Video Generation via MLLM-Based Guidance
authored
a paper
about 3 hours ago
MagicComp: Training-free Dual-Phase Refinement for Compositional Video
Generation
authored
a paper
about 3 hours ago
OpenS2V-Nexus: A Detailed Benchmark and Million-Scale Dataset for
Subject-to-Video Generation
Organizations
upvoted
2
papers
1 day ago