PyraTok: Language-Aligned Pyramidal Tokenizer for Video Understanding and Generation
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866918350100103168 |
|---|---|
| author | Susladkar, Onkar Prakash, Tushar Juvekar, Adheesh Nguyen, Kiet A. Jang, Dong-Hwan Dhillon, Inderjit S Lourentzou, Ismini |
| author_facet | Susladkar, Onkar Prakash, Tushar Juvekar, Adheesh Nguyen, Kiet A. Jang, Dong-Hwan Dhillon, Inderjit S Lourentzou, Ismini |
| contents | Discrete video VAEs underpin modern text-to-video generation and video understanding systems, yet existing tokenizers typically learn visual codebooks at a single scale with limited vocabularies and shallow language supervision, leading to poor cross-modal alignment and zero-shot transfer. We introduce PyraTok, a language-aligned pyramidal tokenizer that learns semantically structured discrete latents across multiple spatiotemporal resolutions. PyraTok builds on a pretrained video VAE and a novel Language aligned Pyramidal Quantization (LaPQ) module that discretizes encoder features at several depths using a shared large binary codebook, yielding compact yet expressive video token sequences. To tightly couple visual tokens with language, PyraTok jointly optimizes multi-scale text-guided quantization and a global autoregressive objective over the token hierarchy. Across ten benchmarks, PyraTok delivers state-of-the-art (SOTA) video reconstruction, consistently improves text-to-video quality, and sets new SOTA zero-shot performance on video segmentation, temporal action localization, and video understanding, scaling robustly to up to 4K/8K resolutions. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2601_16210 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | PyraTok: Language-Aligned Pyramidal Tokenizer for Video Understanding and Generation Susladkar, Onkar Prakash, Tushar Juvekar, Adheesh Nguyen, Kiet A. Jang, Dong-Hwan Dhillon, Inderjit S Lourentzou, Ismini Computer Vision and Pattern Recognition Artificial Intelligence Discrete video VAEs underpin modern text-to-video generation and video understanding systems, yet existing tokenizers typically learn visual codebooks at a single scale with limited vocabularies and shallow language supervision, leading to poor cross-modal alignment and zero-shot transfer. We introduce PyraTok, a language-aligned pyramidal tokenizer that learns semantically structured discrete latents across multiple spatiotemporal resolutions. PyraTok builds on a pretrained video VAE and a novel Language aligned Pyramidal Quantization (LaPQ) module that discretizes encoder features at several depths using a shared large binary codebook, yielding compact yet expressive video token sequences. To tightly couple visual tokens with language, PyraTok jointly optimizes multi-scale text-guided quantization and a global autoregressive objective over the token hierarchy. Across ten benchmarks, PyraTok delivers state-of-the-art (SOTA) video reconstruction, consistently improves text-to-video quality, and sets new SOTA zero-shot performance on video segmentation, temporal action localization, and video understanding, scaling robustly to up to 4K/8K resolutions. |
| title | PyraTok: Language-Aligned Pyramidal Tokenizer for Video Understanding and Generation |
| topic | Computer Vision and Pattern Recognition Artificial Intelligence |
| url | https://arxiv.org/abs/2601.16210 |