6 citations · 15 across the 18 of their papers we have counts for
18 papers
StableVQ: Practical Guidelines for Stable Vector-Quantized Tokenizer Training
Bao Tang, Jiahao Guo, Haoxiang Cao +4
Vector Quantization (VQ) is fundamental to discrete visual tokenizers that power modern autoregressive and masked image generation models. While recent shared-projection codebook m…
Dior: Drawing the Light of Image via Material-Decoupled Illumination Representation
Xuanpu Zhang, Xuesong Niu, Haoxiang Cao +4
Controllable image relighting is an important problem in image editing, and hand-drawn scribbles provide an intuitive interface for specifying the desired illumination. However, ex…
DARS: Dual-Level Credit Assignment RL with Structured Reasoning for Instruction-Based Image Editing
Haoxiang Cao, Jiajiong Cao, Xuanpu Zhang +2
Instruction-based image editing uses a planner-renderer pipeline: a vision-language model (VLM) first converts the instruction into an edit plan, and a diffusion model then execute…
AdvFD: Boosting Visual Generation via Adversarial Fr'echet Distance Loss
Mingju Gao, Jingkai Zhou, Kun Gai +2
Fréchet distance has recently emerged as an effective distribution-level objective for generator post-training, complementing the conventional sample-level diffusion and flow-match…
Where, What, Why, and Importance: Structured Defect Grounding for Text-to-Image Feedback
Huaisong Zhang, Hao Yu, Yuxuan Zhang +7
Despite generating increasingly photorealistic images, text-to-image (T2I) models still exhibit localized, subtle, and structurally complex failures. Diagnosing these failures requ…
MaskAlign: Token-Subset Representation Alignment for Efficient Diffusion Training
Lianyu Pang, Tianlin Pan, Cheng Da +5
Representation alignment with pretrained vision models has recently shown strong potential for accelerating diffusion transformer training. By aligning intermediate diffusion featu…
NormGuard: Reward-Preserving Norm Constraints in Flow-Matching Reinforcement Learning
Tianlin Pan, Lianyu Pang, Cheng Da +4
Reinforcement learning (RL) post-training improves the reward alignment of flow-based generators, but often degrades perceptual quality in ways that are not captured by the reward…
RewardHarness: Self-Evolving Agentic Post-Training
Yuxuan Zhang, Penghui Du, Bo Li +11
Aligning generative models with human preferences remains a central challenge. While RLHF is widely adopted, it relies on large-scale preference data. This limitation is particular…
Principled RL for Flow Matching Emerges from the Chunk-level Policy Optimization
Yifu Luo, Haoyuan Sun, Xinhao Hu +12
Recent Progress in post-training flow matching for text-to-image (T2I) generation with Group Relative Policy Optimization (GRPO) has demonstrated strong potential. However, it is h…
SkyReels-A1: Expressive Portrait Animation in Video Diffusion Transformers
Di Qiu, Zhengcong Fei, Rui Wang +5
We present SkyReels-A1, a simple yet effective framework built upon video diffusion Transformer to facilitate portrait image animation. Existing methodologies still encounter issue…
Ingredients: Blending Custom Photos with Video Diffusion Transformers
Zhengcong Fei, Debang Li, Di Qiu +2
This paper presents a powerful framework to customize video creations by incorporating multiple specific identity (ID) photos, with video diffusion Transformers, referred to as Ing…
Video Diffusion Transformers are In-Context Learners
Zhengcong Fei, Di Qiu, Debang Li +2
This paper investigates a solution for enabling in-context capabilities of video diffusion transformers, with minimal tuning required for activation. Specifically, we propose a sim…
MovieCharacter: A Tuning-Free Framework for Controllable Character Video Synthesis
Di Qiu, Zheng Chen, Rui Wang +4
Recent advancements in character video synthesis still depend on extensive fine-tuning or complex 3D modeling processes, which can restrict accessibility and hinder real-time appli…
FLUX that Plays Music
Zhengcong Fei, Mingyuan Fan, Changqian Yu +1
This paper explores a simple extension of diffusion-based rectified flow Transformers for text-to-music generation, termed as FluxMusic. Generally, along with design in advanced Fl…
Scaling Diffusion Transformers to 16 Billion Parameters
Zhengcong Fei, Mingyuan Fan, Changqian Yu +2
In this paper, we present DiT-MoE, a sparse version of the diffusion Transformer, that is scalable and competitive with dense networks while exhibiting highly optimized inference.…
Dimba: Transformer-Mamba Diffusion Models
Zhengcong Fei, Mingyuan Fan, Changqian Yu +3
This paper unveils Dimba, a new text-to-image diffusion model that employs a distinctive hybrid architecture combining Transformer and Mamba elements. Specifically, Dimba sequentia…
Diffusion-RWKV: Scaling RWKV-Like Architectures for Diffusion Models
Zhengcong Fei, Mingyuan Fan, Changqian Yu +2
Transformers have catalyzed advancements in computer vision and natural language processing (NLP) fields. However, substantial computational complexity poses limitations for their…
Scalable Diffusion Models with State Space Backbone
Zhengcong Fei, Mingyuan Fan, Changqian Yu +1
This paper presents a new exploration into a category of diffusion models built upon state space architecture. We endeavor to train diffusion models for image data, wherein the tra…