Image Diffusion Models Exhibit Emergent Temporal Propagation in Videos • Hacker News Board

Re: Image Diffusion Models Exhibit Emergent Temporal Propagation in Videos

#21

Post by ttul » Thu, Nov 27, 2025, 2:23 PM UTC

ethmarks wrote: post #18

ttul wrote: post #12

This is a cool result. Deep learning image models are trained on enormous amounts of data and the information recorded in their weights continues to astonish me. Over in the Stable Diffusion space, hobbyists (as opposed to professional researchers) are continuing to find new ways to squeeze intelligence out of models that were trained in 2022 and are considerably out of date compared with the latest “flow matching” m…

Gemini 3 is a 10 trillion parameter model?

I read that the pre-training model behind Gemini 3 has 10T parameters. That does not mean that the model they’re serving each day has 10T parameters. The online model is likely distilled from 10T down to something smaller, but I have not had either fact confirmed by Google. These are anecdotes.

Re: Image Diffusion Models Exhibit Emergent Temporal Propagation in Videos

#22

Post by ttul » Thu, Nov 27, 2025, 2:25 PM UTC

cheald wrote: post #20

throwaway314155 wrote: post #19

Earlier quoted context omitted.

I’d love to follow your work. Got a GitHub?

I do (same username), but I haven't published any of this (and in fact my Github has sadly languished lately); I keep working on it with the intent to publish eventually. The big problem with models like this is that the training dynamics have so many degrees of freedom that every time I get close to something I want to publish I end up chasing down another set of rabbit holes. https://gist.github.com/cheald/7d9a436b…

This is super cool work. I’ve built some new sampling techniques for flow matching models that encourage the model to take a “second look” by rewinding sampling to a midpoint and then running the clock forward again. This worked really well with diffusion models (pre-DiT models like SDXL) and I was curious whether it would work with flow matching models like Qwen Image. Yes, it does, but the design is different because flow matching models aren’t de-noising pixels so much as they are simply following a vector field at each step like a ship being pushed by the wind.

Re: Image Diffusion Models Exhibit Emergent Temporal Propagation in Videos

#23

Post by cheald » Thu, Nov 27, 2025, 4:17 PM UTC

ttul wrote: post #22

cheald wrote: post #20

Earlier quoted context omitted.

I do (same username), but I haven't published any of this (and in fact my Github has sadly languished lately); I keep working on it with the intent to publish eventually. The big problem with models like this is that the training dynamics have so many degrees of freedom that every time I get close to something I want to publish I end up chasing down another set of rabbit holes. https://gist.github.com/cheald/7d9a436b…

This is super cool work. I’ve built some new sampling techniques for flow matching models that encourage the model to take a “second look” by rewinding sampling to a midpoint and then running the clock forward again. This worked really well with diffusion models (pre-DiT models like SDXL) and I was curious whether it would work with flow matching models like Qwen Image. Yes, it does, but the design is different becau…

Neat! Is that published anywhere?

It seems conceptually related to ddpm/ancestral sampling, no? Except they're just adding noise to the intermediate latent to simulate a "trajectory jump". How does your method compare?