Level-of-Token Diffusion
cs.CV, cs.AI
Submitted: 2026-10-05
Updated: 2026-10-05
Code: https://github.com/ideogram-oss/ideogram4
Project page: https://georgenakayama.github.io/lotdiffusion
Sources
- Longformer: The Long-Document Transformer
- SAM 3: Segment Anything with Concepts
- Foveated Diffusion: Efficient Spatially Adaptive Image and Video Generation
- Asymmetric Flow Models
- Rethinking Attention with Performers
- Adaptive Length Image Tokenization via Recurrent Allocation
- Vchitect-2.0: Parallel Transformer for Scaling Up Video Diffusion Models
- DC-DiT: Adaptive Compute and Elastic Inference for Visual Generation via Dynamic Chunking
- LoRA: Low-Rank Adaptation of Large Language Models
- DDiT: Dynamic Patch Scheduling for Efficient Diffusion Transformers
- Depth Anything 3: Recovering the Visual Space from Any Views
- FreqCa: Accelerating Diffusion Models via Frequency-Aware Caching
- ToMA: Token Merge with Attention for Diffusion Models
- Latent Consistency Models: Synthesizing High-Resolution Images with Few-Step Inference
- CAT: Content-Adaptive Image Tokenization
- Denoising Diffusion Implicit Models
- Wan: Open and Advanced Large-Scale Video Generative Models
- Linformer: Self-Attention with Linear Complexity
- Human Preference Score v2: A Solid Benchmark for Evaluating Human Preferences of Text-to-Image Synthesis
- Spectral Progressive Diffusion for Efficient Image and Video 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