MambaVideo for Discrete Video Tokenization with Channel-Split Quantization
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866915373465468928 |
|---|---|
| author | Argaw, Dawit Mureja Liu, Xian Chung, Joon Son Liu, Ming-Yu Reda, Fitsum |
| author_facet | Argaw, Dawit Mureja Liu, Xian Chung, Joon Son Liu, Ming-Yu Reda, Fitsum |
| contents | Discrete video tokenization is essential for efficient autoregressive generative modeling due to the high dimensionality of video data. This work introduces a state-of-the-art discrete video tokenizer with two key contributions. First, we propose a novel Mamba-based encoder-decoder architecture that overcomes the limitations of previous sequencebased tokenizers. Second, we introduce a new quantization scheme, channel-split quantization, which significantly enhances the representational power of quantized latents while preserving the token count. Our model sets a new state-of-the-art, outperforming both causal 3D convolutionbased and Transformer-based approaches across multiple datasets. Experimental results further demonstrate its robustness as a tokenizer for autoregressive video generation. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2507_04559 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | MambaVideo for Discrete Video Tokenization with Channel-Split Quantization Argaw, Dawit Mureja Liu, Xian Chung, Joon Son Liu, Ming-Yu Reda, Fitsum Computer Vision and Pattern Recognition Discrete video tokenization is essential for efficient autoregressive generative modeling due to the high dimensionality of video data. This work introduces a state-of-the-art discrete video tokenizer with two key contributions. First, we propose a novel Mamba-based encoder-decoder architecture that overcomes the limitations of previous sequencebased tokenizers. Second, we introduce a new quantization scheme, channel-split quantization, which significantly enhances the representational power of quantized latents while preserving the token count. Our model sets a new state-of-the-art, outperforming both causal 3D convolutionbased and Transformer-based approaches across multiple datasets. Experimental results further demonstrate its robustness as a tokenizer for autoregressive video generation. |
| title | MambaVideo for Discrete Video Tokenization with Channel-Split Quantization |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2507.04559 |