LipsFlower constrains flow matching to stabilize generative image restoration
LipHDiT uses spectral normalization, bounded modulation and finite-radius gain penalties, tested on blind face restoration.
Stories per day · 30 days
2 RSS
Related3d-editing13d-generation1adversarial-robustness1automl1benchmark1cinematic-video1factor-control1flow-matching1geneval1image-restoration1layout1lipschitz1mllm1multimodal1multimodal-ai1open-weights1reference-control1reinforcement-learning1
16 stories, newest first · one card per story
Stories tagged #diffusion-transformer
LLM AutoML searcher jointly tunes architecture, hyperparameters, and data
Trained on one vision task, the searcher reports a 0.91 GenEval score after 896 GPU-days on a billion-parameter Diffusion Transformer.
SeMo3D unifies 3D generation editing and completion
SeMo3D uses selective modeling to unify 3D tasks with region-selective denoising and Omni3DiT architecture.
FAPL trains MLLM latents to control cinematic video generation
A diffusion transformer is grounded on frozen MLLM features, then the model is tuned for sparse prompts and single-factor edits.
MOSS-Image open text-to-image system targets dense text and layouts
A 5B diffusion transformer plus an agentic prompt module reports 96.3% average text coverage; weights are promised for release.
OmniXRef unifies text, image, and video references for editing
A diffusion transformer jointly edits a source video from variable text, image, and video references, and the authors say it beats commercial models.
GRACE: Generation-aware latent compression for efficient video generation
GRACE cuts Wan2.1-I2V-14B video tokens 8x and latency 11.1x without losing VBench quality.
LiFT: Loop Flow Transformers
LiFT loops a shared Diffusion Transformer and lowers ImageNet FID while cutting parameters and FLOPs.
canberkkkkkk/ema-lightning — new model trending #30 on Hugging Face
EMA Lightning is an 8.6M-parameter Apache-2.0 Turkish TTS model reporting 0.92% WER on Freya-TR-Eval.
4
ProtoFlow: Prototype-Guided Flow Matching for Multivariate Time Series Forecasting
ProtoFlow forecasts multivariate time series by flow-matching from learned vector-quantization prototypes instead of Gaussian noise.
Video DeltaNet: A Video-Native Hybrid Attention for Livestream Video Generation
Video DeltaNet combines softmax and linear attention for video generation, cutting DiT denoising of a 14.3-second 768p video to 6.7 seconds.
Video DeltaNet: A Video-Native Hybrid Attention for Livestream Video Generation
Video DeltaNet combines local Softmax attention with linear memory, enabling 14.5x faster 768p video generation on eight NVIDIA B200 GPUs.
LynnReal-Omni: Native multi-modal Video Generation for Agentic Visual Workflows
LynnReal-Omni unifies controllable video generation tasks in a 32B multimodal diffusion transformer, with a 27B Flash variant rendering 540p clips in 377 ms.
StepAudio 3 Music Technical Report
StepAudio 3 Music introduces long-form text-controlled music generation using ABC-notation planning and flow-matching diffusion, ranking near the top music arena.
UniMate: One Unified Model to Animate Diverse Skeletons
Researchers introduce UniMate, a topology-aware diffusion transformer generating text-driven motion for arbitrary 3D skeletons without per-skeleton retraining.
UniMate: One Unified Model to Animate Diverse Skeletons
UniMate is a topology-aware diffusion transformer generating articulated motion for arbitrary rigged skeletons from text, trained on 13,006 motion sequences.