TrajCFG: Token-Wise Classifier-Free Guidance via Denoising Trajectory Feedback
cs.CV
Submitted: 2026-10-09
Updated: 2026-10-09
Comments: 22 pages. Project page: https://jyliu-98.github.io/trajcfg-project-page/
Project page: https://jyliu-98.github.io/trajcfg-project-page
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Sources
- World Simulation with Video Foundation Models for Physical AI
- Steering Video Diffusion Transformers with Massive Activations
- CFG-Zero*: Improved Classifier-Free Guidance for Flow Matching Models
- C$^2$FG: Control Classifier-Free Guidance via Score Discrepancy Analysis
- LTX-2: Efficient Joint Audio-Visual Foundation Model
- Classifier-Free Diffusion Guidance
- ELLA: Equip Diffusion Models with LLM for Enhanced Semantic Alignment
- Delta Score Matters! Spatial Adaptive Multi Guidance in Diffusion Models
- Ovi: Twin Backbone Cross-Modal Fusion for Audio-Video Generation
- VAGS: Velocity Adaptive Guidance Scale for Image Editing and Generation
- PathGuide: Dynamic Classifier-Free Guidance via On-Policy Transport Alignment
- The 2017 DAVIS Challenge on Video Object Segmentation
- Wan: Open and Advanced Large-Scale Video Generative Models
- Analysis of Classifier-Free Guidance Weight Schedulers
- Qwen-Image Technical Report
- SpecLoR: Spectral Lookahead Rectification for Motion-Coherent Text-to-Video Generation
- Hunyuan3D 2.0: Scaling Diffusion Models for High Resolution Textured 3D Assets Generation
Related papers
- Loss Knows Best: Detecting Annotation Errors in Videos via Loss Trajectories
- AnchorWeave: World-Consistent Video Generation with Retrieved Local Spatial Memories
- Benchmarking the Robustness of Foundation Models for Mammography under Domain Shift
- MambaX-Net: Dual-Input Mamba-Enhanced Cross-Attention Network for Longitudinal MRI Segmentation
- TeleOCR: Navigating Document Parsing Across Digital and Camera-Captured Documents
- A Survey on Efficient Vision-Language-Action Models