Search

"classifier-free guidance"8 resultsKeyword

LTX-2: Efficient Joint Audio-Visual Foundation Model

Yoav HaCohen, Benny Brazowski et al.Jan 6arXiv

LTX-2 is an open-source model that makes video and sound together from a text prompt, so the picture and audio match in time and meaning.

#text-to-video#text-to-audio#audiovisual generation

Not triaged yet

RecTok: Reconstruction Distillation along Rectified Flow

Intermediate

Qingyu Shi, Size Wu et al.Dec 15arXiv

RecTok is a new visual tokenizer that teaches the whole training path of a diffusion model (the forward flow) to be smart about image meaning, not just the starting latent features.

#Rectified Flow#Flow Matching#Visual Tokenizer

Not triaged yet

CFG-Ctrl: Control-Based Classifier-Free Diffusion Guidance

Intermediate

Hanyang Wang, Yiyang Liu et al.Mar 3arXiv

This paper turns a popular image-guidance trick (Classifier-Free Guidance) into a feedback-control problem, just like keeping a car steady in its lane.

#Classifier-Free Guidance#Sliding Mode Control#Diffusion Models

Not triaged yet

Semantic Routing: Exploring Multi-Layer LLM Feature Weighting for Diffusion Transformers

Intermediate

Bozhou Li, Yushuo Guan et al.Feb 3arXiv

The paper shows that using information from many layers of a language model (not just one) helps text-to-image diffusion transformers follow prompts much better.

#Diffusion Transformer#Text Conditioning#Multi-layer LLM Features

Not triaged yet

CAPTAIN: Semantic Feature Injection for Memorization Mitigation in Text-to-Image Diffusion Models

Intermediate

Tong Zhang, Carlos Hinojosa et al.Dec 11arXiv

Diffusion models sometimes copy training images too closely, which can be a privacy and copyright problem.

#diffusion models#memorization mitigation#latent feature injection

Not triaged yet

PixelGen: Pixel Diffusion Beats Latent Diffusion with Perceptual Loss

Intermediate

Zehong Ma, Ruihan Xu et al.Feb 2arXiv

PixelGen is a new image generator that works directly with pixels and uses what-looks-good-to-people guidance (perceptual loss) to improve quality.

#pixel diffusion#perceptual loss#LPIPS

Not triaged yet

Boosting Latent Diffusion Models via Disentangled Representation Alignment

Intermediate

John Page, Xuesong Niu et al.Jan 9arXiv

This paper shows that the best VAEs for image generation are the ones whose latents neatly separate object attributes, a property called semantic disentanglement.

#Send-VAE#semantic disentanglement#latent diffusion

Not triaged yet

SARAH: Spatially Aware Real-time Agentic Humans

Intermediate

Evonne Ng, Siwei Zhang et al.Feb 20arXiv

SARAH is a real-time system that makes virtual characters move their whole bodies naturally during a conversation while knowing where the user is.

#spatially aware motion#real-time avatars#causal transformer

Not triaged yet