Accelerated Diffusion Models via Speculative Sampling
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2025
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866915372820594688 |
|---|---|
| author | De Bortoli, Valentin Galashov, Alexandre Gretton, Arthur Doucet, Arnaud |
| author_facet | De Bortoli, Valentin Galashov, Alexandre Gretton, Arthur Doucet, Arnaud |
| contents | Speculative sampling is a popular technique for accelerating inference in Large Language Models by generating candidate tokens using a fast draft model and accepting or rejecting them based on the target model's distribution. While speculative sampling was previously limited to discrete sequences, we extend it to diffusion models, which generate samples via continuous, vector-valued Markov chains. In this context, the target model is a high-quality but computationally expensive diffusion model. We propose various drafting strategies, including a simple and effective approach that does not require training a draft model and is applicable out of the box to any diffusion model. Our experiments demonstrate significant generation speedup on various diffusion models, halving the number of function evaluations, while generating exact samples from the target model. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2501_05370 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Accelerated Diffusion Models via Speculative Sampling De Bortoli, Valentin Galashov, Alexandre Gretton, Arthur Doucet, Arnaud Machine Learning Speculative sampling is a popular technique for accelerating inference in Large Language Models by generating candidate tokens using a fast draft model and accepting or rejecting them based on the target model's distribution. While speculative sampling was previously limited to discrete sequences, we extend it to diffusion models, which generate samples via continuous, vector-valued Markov chains. In this context, the target model is a high-quality but computationally expensive diffusion model. We propose various drafting strategies, including a simple and effective approach that does not require training a draft model and is applicable out of the box to any diffusion model. Our experiments demonstrate significant generation speedup on various diffusion models, halving the number of function evaluations, while generating exact samples from the target model. |
| title | Accelerated Diffusion Models via Speculative Sampling |
| topic | Machine Learning |
| url | https://arxiv.org/abs/2501.05370 |