Purrception: Variational Flow Matching for Vector-Quantized Image Generation
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866917339790835712 |
|---|---|
| author | Matişan, Răzvan-Andrei Hu, Vincent Tao Bartosh, Grigory Ommer, Björn Snoek, Cees G. M. Welling, Max van de Meent, Jan-Willem Derakhshani, Mohammad Mahdi Eijkelboom, Floor |
| author_facet | Matişan, Răzvan-Andrei Hu, Vincent Tao Bartosh, Grigory Ommer, Björn Snoek, Cees G. M. Welling, Max van de Meent, Jan-Willem Derakhshani, Mohammad Mahdi Eijkelboom, Floor |
| contents | We introduce Purrception, a variational flow matching approach for vector-quantized image generation that provides explicit categorical supervision while maintaining continuous transport dynamics. Our method adapts Variational Flow Matching to vector-quantized latents by learning categorical posteriors over codebook indices while computing velocity fields in the continuous embedding space. This combines the geometric awareness of continuous methods with the discrete supervision of categorical approaches, enabling uncertainty quantification over plausible codes and temperature-controlled generation. We evaluate Purrception on ImageNet-1k 256x256 generation. Training converges faster than both continuous flow matching and discrete flow matching baselines while achieving competitive FID scores with state-of-the-art models. This demonstrates that Variational Flow Matching can effectively bridge continuous transport and discrete supervision for improved training efficiency in image generation. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2510_01478 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Purrception: Variational Flow Matching for Vector-Quantized Image Generation Matişan, Răzvan-Andrei Hu, Vincent Tao Bartosh, Grigory Ommer, Björn Snoek, Cees G. M. Welling, Max van de Meent, Jan-Willem Derakhshani, Mohammad Mahdi Eijkelboom, Floor Computer Vision and Pattern Recognition Artificial Intelligence Machine Learning We introduce Purrception, a variational flow matching approach for vector-quantized image generation that provides explicit categorical supervision while maintaining continuous transport dynamics. Our method adapts Variational Flow Matching to vector-quantized latents by learning categorical posteriors over codebook indices while computing velocity fields in the continuous embedding space. This combines the geometric awareness of continuous methods with the discrete supervision of categorical approaches, enabling uncertainty quantification over plausible codes and temperature-controlled generation. We evaluate Purrception on ImageNet-1k 256x256 generation. Training converges faster than both continuous flow matching and discrete flow matching baselines while achieving competitive FID scores with state-of-the-art models. This demonstrates that Variational Flow Matching can effectively bridge continuous transport and discrete supervision for improved training efficiency in image generation. |
| title | Purrception: Variational Flow Matching for Vector-Quantized Image Generation |
| topic | Computer Vision and Pattern Recognition Artificial Intelligence Machine Learning |
| url | https://arxiv.org/abs/2510.01478 |