# Adaptive Tokenisation Via Temporal Redundancy Masking And Latent Inpainting

arXiv: https://arxiv.org/abs/2606.06158 · HTML: https://mlsanity.com/paper/2606.06158/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 31 of its 37 references that are in the map.
- First posted on arXiv 4 Jun 2026.

## Builds on (most cited first, 30 of 31)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://mlsanity.com/paper/2010.11929/) (2020)
- [The Unreasonable Effectiveness of Deep Features as a Perceptual Metric](https://mlsanity.com/paper/1801.03924/) (2018)
- [Neural Discrete Representation Learning](https://mlsanity.com/paper/1711.00937/) (2017)
- [RoFormer: Enhanced Transformer with Rotary Position Embedding](https://mlsanity.com/paper/2104.09864/) (2021)
- [Taming Transformers for High-Resolution Image Synthesis](https://mlsanity.com/paper/2012.09841/) (2020)
- [The Kinetics Human Action Video Dataset](https://mlsanity.com/paper/1705.06950/) (2017)
- [ViViT: A Video Vision Transformer](https://mlsanity.com/paper/2103.15691/) (2021)
- [Align your Latents: High-Resolution Video Synthesis with Latent Diffusion Models](https://mlsanity.com/paper/2304.08818/) (2023)
- [AnimateDiff: Animate Your Personalized Text-to-Image Diffusion Models without Specific Tuning](https://mlsanity.com/paper/2307.04725/) (2023)
- [Towards Accurate Generative Models of Video: A New Metric & Challenges](https://mlsanity.com/paper/1812.01717/) (2018)
- [WorldSimBench: Towards Video Generation Models as World Simulators](https://mlsanity.com/paper/2410.18072/) (2024)
- [Cosmos World Foundation Model Platform for Physical AI](https://mlsanity.com/paper/2501.03575/) (2025)
- [VideoGPT: Video Generation using VQ-VAE and Transformers](https://mlsanity.com/paper/2104.10157/) (2021)
- [Language Model Beats Diffusion -- Tokenizer is Key to Visual Generation](https://mlsanity.com/paper/2310.05737/) (2023)
- [Revisiting Feature Prediction for Learning Visual Representations from Video](https://mlsanity.com/paper/2404.08471/) (2024)
- [Sora: A Review on Background, Technology, Limitations, and Opportunities of Large Vision Models](https://mlsanity.com/paper/2402.17177/) (2024)
- [MAGVIT: Masked Generative Video Transformer](https://mlsanity.com/paper/2212.05199/) (2022)
- [Autoregressive Video Generation without Vector Quantization](https://mlsanity.com/paper/2412.14169/) (2024)
- [OmniTokenizer: A Joint Image-Video Tokenizer for Visual Generation](https://mlsanity.com/paper/2406.09399/) (2024)
- [The 2019 DAVIS Challenge on VOS: Unsupervised Multi-Object Segmentation](https://mlsanity.com/paper/1905.00737/) (2019)
- [CV-VAE: A Compatible Video VAE for Latent Generative Video Models](https://mlsanity.com/paper/2405.20279/) (2024)
- [Matryoshka Query Transformer for Large Vision-Language Models](https://mlsanity.com/paper/2405.19315/) (2024)
- [ElasticTok: Adaptive Tokenization for Image and Video](https://mlsanity.com/paper/2410.08368/) (2024)
- [LARP: Tokenizing Videos with a Learned Autoregressive Generative Prior](https://mlsanity.com/paper/2410.21264/) (2024)
- [Don't Look Twice: Faster Video Transformers with Run-Length Tokenization](https://mlsanity.com/paper/2411.05222/) (2024)
- [Learning Adaptive and Temporally Causal Video Tokenization in a 1D Latent Space](https://mlsanity.com/paper/2505.17011/) (2025)
- [Single-pass Adaptive Image Tokenization for Minimum Program Search](https://mlsanity.com/paper/2507.07995/) (2025)
- [InfoTok: Adaptive Discrete Video Tokenizer via Information-Theoretic Compression](https://mlsanity.com/paper/2512.16975/) (2025)
- [Robust Vector Quantized-Variational Autoencoder](https://mlsanity.com/paper/2202.01987/) (2022)

## Abstract (arXiv)

Adaptive video tokenisation seeks to dynamically allocate token budgets based on the underlying visual complexity of a sequence. Current continuous-regime approaches achieve this via iterative binarised searches or trained neural regressors, while discrete methods often require a full-rate decoder pass to estimate information content. We demonstrate that such computational overheads are not strictly necessary. We show that the latent space of a frozen continuous video tokeniser inherently encodes temporal redundancy that can be exploited directly: spatial positions whose latent representations change minimally between consecutive frames carry near-zero additional information. We introduce a parameter-free adaptive token allocation mechanism that applies a fixed threshold to per-position temporal-L1 differences, identifying and dropping redundant latent positions. Consequently, the compression rate emerges naturally from the input content rather than being enforced top-down: static scenes get compressed aggressively, while highly dynamic sequences retain more tokens. To reconstruct the dropped positions, we propose the Latent Inpainting Transformer (LIT), a lightweight factorised spatial-temporal attention architecture. The resulting inference pipeline is highly efficient, requiring only a single encoder pass and one LIT forward pass, eliminating the need for auxiliary routing networks. Evaluations across TokenBench and DAVIS, which are the standard benchmarks used by recent tokenisers~citeinfotok, agarwal2025cosmos, indicate that our framework yields meaningful, content-driven token allocation while maintaining competitive reconstruction fidelity, and delivers a 31× inference-time speedup over the continuous adaptive baseline (ElasticTok-CV) and an ≈2× speedup over the discrete information-theoretic baseline (InfoTok).