VMix: Improving Text-to-Image Diffusion Model with Cross-Attention Mixing Control
Paper • 2412.20800 • Published • 11
Generation Quality Enhancement
-
-
Padding Tone: A Mechanistic Analysis of Padding Tokens in T2I Models
Paper • 2501.06751 • Published • 32 -
Inference-Time Scaling for Diffusion Models beyond Scaling Denoising Steps
Paper • 2501.09732 • Published • 72 -
Learnings from Scaling Visual Tokenizers for Reconstruction and Generation
Paper • 2501.09755 • Published • 35 -
Diffusion-Sharpening: Fine-tuning Diffusion Models with Denoising Trajectory Sharpening
Paper • 2502.12146 • Published • 16 -
PLADIS: Pushing the Limits of Attention in Diffusion Models at Inference Time by Leveraging Sparsity
Paper • 2503.07677 • Published • 86 -
CFG-Zero*: Improved Classifier-Free Guidance for Flow Matching Models
Paper • 2503.18886 • Published • 24 -
Alchemist: Turning Public Text-to-Image Data into Generative Gold
Paper • 2505.19297 • Published • 84 -
Rethinking Cross-Modal Interaction in Multimodal Diffusion Transformers
Paper • 2506.07986 • Published • 19 -
Ambient Diffusion Omni: Training Good Models with Bad Data
Paper • 2506.10038 • Published • 9 -
Pref-GRPO: Pairwise Preference Reward-based GRPO for Stable Text-to-Image Reinforcement Learning
Paper • 2508.20751 • Published • 90 -
Image Tokenizer Needs Post-Training
Paper • 2509.12474 • Published • 9 -
One Small Step in Latent, One Giant Leap for Pixels: Fast Latent Upscale Adapter for Your Diffusion Models
Paper • 2511.10629 • Published • 129 -
The Image as Its Own Reward: Reinforcement Learning with Adversarial Reward for Image Generation
Paper • 2511.20256 • Published • 28 -
TDM-R1: Reinforcing Few-Step Diffusion Models with Non-Differentiable Reward
Paper • 2603.07700 • Published • 13