FlexTok: Resampling Images into 1D Token Sequences of Flexible Length
Roman Bachmann, Jesse Allardice, David Mizrahi, Enrico Fini, Oguzhan Fatih Kar, Elmira Amirloo, Alaaeldin El-Nouby, Amir Zamir, Afshin Dehghan
Abstract
FlexTok: Resampling Images into 1D Token Sequences of Flexible Length 256 tokens 128 tokens 64 tokens 32 tokens 16 tokens 8 tokens 4 tokens 2 tokens 1 token Original RGB Figure 2: Reconstruction examples using FlexTok d18-d28 trained on DFN. Notice how most of the images' semantic and geometric content is captured by fewer than 16 tokens. The first tokens already capture the high-level semantic concepts (e.g., gray bird, people in colorful garments, mountain scene, yellow flower), while more tokens are required to reconstruct more intricate scene details (e.g., position and clothing of every person, brushstroke placement, etc.). To showcase out-of-distribution reconstruction, we generated the original images using Midjourney v6.1 (Midjourney, 2024). * Note. For evaluation of the class-conditioned image generation results, we follow the common practice of measuring the generation FID (gFID) of 50K generated samples relative to the reference statistics calculated over the entire training split of the ImageNet-1k dataset (Dhariwal & Nichol, 2021).
Ask Lune
Know this paper in minutes, not an afternoon.
Lune can read every page of it and the top-tier papers that cite it, and gives you its claims, its evidence and its limits, each one cited to the paper it comes from. No skimming, no AI slop.
54 top-tier papers have cited it since. Know what they found before your reviewers do.
Lune can draw on
the full paper · the top-tier papers that cite it · related work
Free to start. No credit card required.
Lineage
- High-Resolution Image Synthesis with Latent Diffusion Models cites Diffusion Models Beat GANs on Image Synthesis
- High-Resolution Image Synthesis with Latent Diffusion Models cites Zero-Shot Text-to-Image Generation
- Captain Cinema: Towards Short Movie Generation cites High-Resolution Image Synthesis with Latent Diffusion Models
- Captain Cinema: Towards Short Movie Generation cites Scalable Diffusion Models with Transformers
- Scalable Diffusion Models with Transformers cites High-Resolution Image Synthesis with Latent Diffusion Models
- Scalable Diffusion Models with Transformers cites Diffusion Models Beat GANs on Image Synthesis
- Scalable Diffusion Models with Transformers cites Zero-Shot Text-to-Image Generation
- Diffusion Models Beat GANs on Image Synthesis cites Zero-Shot Text-to-Image Generation
- Latent Denoising Makes Good Tokenizers cites High-Resolution Image Synthesis with Latent Diffusion Models
- Latent Denoising Makes Good Tokenizers cites Scalable Diffusion Models with Transformers
- Latent Denoising Makes Good Tokenizers cites Diffusion Models Beat GANs on Image Synthesis
- Latent Denoising Makes Good Tokenizers cites Scaling Rectified Flow Transformers for High-Resolution Image Synthesis
- Lost in Latent Space: An Empirical Study of Latent Diffusion Models for Physics Emulation cites High-Resolution Image Synthesis with Latent Diffusion Models
- Lost in Latent Space: An Empirical Study of Latent Diffusion Models for Physics Emulation cites Scalable Diffusion Models with Transformers
- Lost in Latent Space: An Empirical Study of Latent Diffusion Models for Physics Emulation cites Scaling Rectified Flow Transformers for High-Resolution Image Synthesis
- Scaling Rectified Flow Transformers for High-Resolution Image Synthesis cites High-Resolution Image Synthesis with Latent Diffusion Models
- Scaling Rectified Flow Transformers for High-Resolution Image Synthesis cites Scalable Diffusion Models with Transformers
- Scaling Rectified Flow Transformers for High-Resolution Image Synthesis cites Diffusion Models Beat GANs on Image Synthesis
- Vision Foundation Models as Effective Visual Tokenizers for Autoregressive Generation cites Scalable Diffusion Models with Transformers
Cited by top-tier papers54
- Frame Context Packing and Drift Prevention in Next-Frame-Prediction Video Diffusion ModelsLvmin Zhang, Shengqu Cai, Muyang Li, Gordon Wetzstein et al.NeurIPS 2025 · 132 citations
- Captain Cinema: Towards Short Movie GenerationJunfei Xiao, Ceyuan Yang, Lvmin Zhang, Shengqu Cai et al.ICLR 2026 · 51 citations
- Lost in Latent Space: An Empirical Study of Latent Diffusion Models for Physics EmulationFrançois Rozet, Ruben Ohana, Michael McCabe, Gilles Louppe et al.NeurIPS 2025 · 23 citations
- Vision Foundation Models as Effective Visual Tokenizers for Autoregressive GenerationAnlin Zheng, Xin Wen, Xuanyang Zhang, Chuofan Ma et al.NeurIPS 2025 · 21 citations
- Latent Denoising Makes Good TokenizersJiawei Yang, Tianhong Li, Lijie Fan, Yonglong Tian et al.ICLR 2026 · 17 citations
Builds on36
- Diffusion Models Beat GANs on Image SynthesisPrafulla Dhariwal, Alexander Quinn NicholNeurIPS 2021 · 13,211 citations
- High-Resolution Image Synthesis with Latent Diffusion ModelsRobin Rombach, Andreas Blattmann, Dominik Lorenz, Patrick Esser et al.CVPR 2022 · 13,123 citations
- Zero-Shot Text-to-Image GenerationAditya Ramesh, Mikhail Pavlov, Gabriel Goh, Scott Gray et al.ICML 2021 · 6,356 citations
- Scalable Diffusion Models with TransformersWilliam Peebles, Saining XieICCV 2023 · 5,568 citations
- Scaling Rectified Flow Transformers for High-Resolution Image SynthesisPatrick Esser, Sumith Kulal, Andreas Blattmann, Rahim Entezari et al.ICML 2024 · 3,620 citations
Related papers
- VideoFlexTok: Flexible-Length Coarse-to-Fine Video TokenizationAndrei Atanov, Jesse Allardice, Roman Bachmann, Oğuzhan Fatih Kar et al.ICML 2026 · 3 citations
- An Image is Worth 32 Tokens for Reconstruction and GenerationQihang Yu, Mark Weber, Xueqing Deng, Xiaohui Shen et al.NeurIPS 2024 · 331 citations
- Prompt Yourself: Awakening Textual Semantics in 1D Visual TokenizersHualiang Wang, Siming Fu, Weinan Jia, Yuning Lu et al.CVPR 2026
- Language-Guided Image Tokenization for GenerationKaiwen Zha, Lijun Yu, Alireza Fathi, David A. Ross et al.CVPR 2025
- Towards Sequence Modeling Alignment between Tokenizer and Autoregressive ModelPingyu Wu, Kai Zhu, Yu Liu, Longxiang Tang et al.ICLR 2026 · 16 citations
