DiverseAR: Boosting Diversity in Bitwise Autoregressive Image Generation
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_ | 1866909940404191232 |
|---|---|
| author | Yang, Ying Lv, Zhengyao Pan, Tianlin Wang, Haofan Yang, Binxin Yin, Hubery Li, Chen Si, Chenyang |
| author_facet | Yang, Ying Lv, Zhengyao Pan, Tianlin Wang, Haofan Yang, Binxin Yin, Hubery Li, Chen Si, Chenyang |
| contents | In this paper, we investigate the underexplored challenge of sample diversity in autoregressive (AR) generative models with bitwise visual tokenizers. We first analyze the factors that limit diversity in bitwise AR models and identify two key issues: (1) the binary classification nature of bitwise modeling, which restricts the prediction space, and (2) the overly sharp logits distribution, which causes sampling collapse and reduces diversity. Building on these insights, we propose DiverseAR, a principled and effective method that enhances image diversity without sacrificing visual quality. Specifically, we introduce an adaptive logits distribution scaling mechanism that dynamically adjusts the sharpness of the binary output distribution during sampling, resulting in smoother predictions and greater diversity. To mitigate potential fidelity loss caused by distribution smoothing, we further develop an energy-based generation path search algorithm that avoids sampling low-confidence tokens, thereby preserving high visual quality. Extensive experiments demonstrate that DiverseAR substantially improves sample diversity in bitwise autoregressive image generation. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2512_02931 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | DiverseAR: Boosting Diversity in Bitwise Autoregressive Image Generation Yang, Ying Lv, Zhengyao Pan, Tianlin Wang, Haofan Yang, Binxin Yin, Hubery Li, Chen Si, Chenyang Computer Vision and Pattern Recognition In this paper, we investigate the underexplored challenge of sample diversity in autoregressive (AR) generative models with bitwise visual tokenizers. We first analyze the factors that limit diversity in bitwise AR models and identify two key issues: (1) the binary classification nature of bitwise modeling, which restricts the prediction space, and (2) the overly sharp logits distribution, which causes sampling collapse and reduces diversity. Building on these insights, we propose DiverseAR, a principled and effective method that enhances image diversity without sacrificing visual quality. Specifically, we introduce an adaptive logits distribution scaling mechanism that dynamically adjusts the sharpness of the binary output distribution during sampling, resulting in smoother predictions and greater diversity. To mitigate potential fidelity loss caused by distribution smoothing, we further develop an energy-based generation path search algorithm that avoids sampling low-confidence tokens, thereby preserving high visual quality. Extensive experiments demonstrate that DiverseAR substantially improves sample diversity in bitwise autoregressive image generation. |
| title | DiverseAR: Boosting Diversity in Bitwise Autoregressive Image Generation |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2512.02931 |