Upsample What Matters: Region-Adaptive Latent Sampling for Accelerated Diffusion Transformers Paper • 2507.08422 • Published Jul 11 • 35
view article Article Announcing NVIDIA Cosmos World Foundation Models By mingyuliutw and 1 other • Jan 7 • 26
Efficient Personalization of Quantized Diffusion Model without Backpropagation Paper • 2503.14868 • Published Mar 19 • 20
PLADIS: Pushing the Limits of Attention in Diffusion Models at Inference Time by Leveraging Sparsity Paper • 2503.07677 • Published Mar 10 • 87
Skrr: Skip and Re-use Text Encoder Layers for Memory Efficient Text-to-Image Generation Paper • 2502.08690 • Published Feb 12 • 44
INTRA: Interaction Relationship-aware Weakly Supervised Affordance Grounding Paper • 2409.06210 • Published Sep 10, 2024 • 27
BeyondScene: Higher-Resolution Human-Centric Scene Generation With Pretrained Diffusion Paper • 2404.04544 • Published Apr 6, 2024 • 24