Rethinking Cross-Layer Information Routing in Diffusion Transformers
Paper • 2605.20708 • Published • 94
AI & ML interests
None defined yet.
Recent Activity
View all activity
Papers
Rethinking Cross-Layer Information Routing in Diffusion Transformers
Full Attention Strikes Back: Transferring Full Attention into Sparse within Hundred Training Steps
zykRichard
authored a
paper about 12 hours ago
Met4physics
submitted a
paper to Daily Papers 1 day ago
zykRichard
submitted a
paper to Daily Papers 4 days ago
zykRichard
authored a
paper 5 days ago
zykRichard
updated a
model 5 months ago
DuoAttention(ICLR 2025)
1
#1 opened 5 months ago
by 
11agentone
smy111
updated a
model 5 months ago
smy111
published a
model 5 months ago