STARFlow2: Bridging Language Models and Normalizing Flows for Unified Multimodal Generation
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , , , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2026
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866914545388224512 |
|---|---|
| author | Shen, Ying Chen, Tianrong Gao, Yuan Zhang, Yizhe Wang, Yuyang Bautista, Miguel Ángel Zhai, Shuangfei Susskind, Joshua M. Gu, Jiatao |
| author_facet | Shen, Ying Chen, Tianrong Gao, Yuan Zhang, Yizhe Wang, Yuyang Bautista, Miguel Ángel Zhai, Shuangfei Susskind, Joshua M. Gu, Jiatao |
| contents | Deep generative models have advanced rapidly across text and vision, motivating unified multimodal systems that can understand, reason over, and generate interleaved text-image sequences. Most existing approaches combine autoregressive language modeling with diffusion-based image generators, inheriting a structural mismatch between causal text generation and iterative visual denoising. We observe that autoregressive normalizing flows are autoregressive Transformers--sharing the same causal mask, KV-cache mechanism, and left-to-right structure as LLMs--making them the most natural paradigm for true unified multimodal generation. We present STARFlow2, built on the Pretzel architecture that vertically interleaves a pretrained VLM stream with a TarFlow stream via residual skip connections, both operating under the same causal mask. Combined with a deep-shallow flow design and a unified FAE latent space, STARFlow2 enables cache-friendly interleaved generation where both text and visual outputs directly enter the KV-cache without re-encoding. Experiments demonstrate strong performance across image generation and multimodal understanding benchmarks, validating autoregressive flows as a viable foundation for unified multimodal modeling. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2605_08029 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | STARFlow2: Bridging Language Models and Normalizing Flows for Unified Multimodal Generation Shen, Ying Chen, Tianrong Gao, Yuan Zhang, Yizhe Wang, Yuyang Bautista, Miguel Ángel Zhai, Shuangfei Susskind, Joshua M. Gu, Jiatao Computer Vision and Pattern Recognition Machine Learning Deep generative models have advanced rapidly across text and vision, motivating unified multimodal systems that can understand, reason over, and generate interleaved text-image sequences. Most existing approaches combine autoregressive language modeling with diffusion-based image generators, inheriting a structural mismatch between causal text generation and iterative visual denoising. We observe that autoregressive normalizing flows are autoregressive Transformers--sharing the same causal mask, KV-cache mechanism, and left-to-right structure as LLMs--making them the most natural paradigm for true unified multimodal generation. We present STARFlow2, built on the Pretzel architecture that vertically interleaves a pretrained VLM stream with a TarFlow stream via residual skip connections, both operating under the same causal mask. Combined with a deep-shallow flow design and a unified FAE latent space, STARFlow2 enables cache-friendly interleaved generation where both text and visual outputs directly enter the KV-cache without re-encoding. Experiments demonstrate strong performance across image generation and multimodal understanding benchmarks, validating autoregressive flows as a viable foundation for unified multimodal modeling. |
| title | STARFlow2: Bridging Language Models and Normalizing Flows for Unified Multimodal Generation |
| topic | Computer Vision and Pattern Recognition Machine Learning |
| url | https://arxiv.org/abs/2605.08029 |