LoGG next paper: Accelerated Diffusion Models via Speculative Sampling

[Hannes Stärk]

Hi together,

Next paper: 

Accelerated Diffusion Models via Speculative Sampling https://arxiv.org/abs/2501.05370 (Valentin De Bortoli, Alexandre Galashov, Arthur Gretton, Arnaud Doucet)
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.

Speaker:
The myth, the legend, Valentin De Bortoli from Google Deepmind.

Meeting Details:
Every Monday at 12:00 ET / 9:00 PT / 18:00 CE(S)T.  
https://zoom.us/j/5775722530?pwd=ZzlGTXlDNThhUDZOdU4vN2JRMm5pQT09

Add it to your calendar:
Subscribe via Google Calendar, or subscribe via iCal.
Alternatively, add the events, or add this single event.

Slack Workspace for discussion and paper voting:
https://join.slack.com/t/logag/shared_invite/zt-2nmy5vsj7-8KEDXe~mU~p5fCxwupSv5w

All information: Schedule of upcoming papers, recordings, etc.:
https://portal.valencelabs.com/logg