High-resolution image generation (4-16 megapixels) on memory-constrained mobile devices requires patch-based latent space processing rather than scaling the entire latent space proportionally to output resolution.
PixelRush enables 4-16 megapixel generation on mobile devices by generating a base 1K image, upsampling it, then processing patches independently in latent space with guided semantic noise, achieving 35x speedup (10 minutes to 20 seconds). ✦ AI generated
Fatih Porikli · The TWIML AI Podcast · 2026-08-12 · original ↗
starts at this moment · 33:49
“what's different about the way you've approached this?”
we are saying that can we even push it to the next level because before we have been talking about let's say 1K resolution... but the question is now can we do like 4 megapixel image generation, 16 megapixel image generation. The challenge is not only how fast you can run the model... but there's a memory challenge also because when the image resolution gets larger we need to retain this diffusion process the latent features in the memory somewhere on the device... we partition the input image into like original size patches let's say I like 1K 1K image. Now I have like 4K 4K... so I divide it into 16 parts... we allow generating noise in the background... it's maybe 35 times faster.
verbatim transcript · starts at 33:49
33:49a different objective. Now we ask how to generate images much more efficiently than even what we did before because Qualcomm proposed and you know presented many papers at the previous CDPRs how to run such models much more efficiently on mobile phones you know but now we are saying that can we even push it to the next level because before we have been talking about let's say 1K resolution
34:15and that's the current sort of even the cloud models are kind of limited to that resolution. But the question is now can we do like uh 4 megapixel image generation, 16 megapixel image generation. The challenge is not only you know kind of like how fast you can run the model. By the way such models existing solutions may take anywhere from you know kind of uh seconds like 50 seconds
34:47to minutes. they are not that fast. Um um but there's a memory challenge also because when the image resolution gets larger we need to retain this diffusion process uh the latent features in the memory somewhere on the device. So that uh requires big footprints. So how can we run a model such that we generate a extremely large image high resolution image and it of course the quality has
35:22to be also still you know real high resolution not just you know kind of like high resolution of sampled image but real details a lot of details are there and it will run at a reasonable time you don't need to wait like 10 minutes which is the current you know kind of uh I mean XFR solution uh the time such model states requires and it would run on let's say a memory
35:48available on a handh held device like a smartphone. Yeah, those are that was our objective uh for the let's say pixel rush and also we didn't want to change really the existing models significantly. So that is something to note. uh there are wonderful uh image generation models and you know some of them kind of like by many big companies kind of we are all we all know uh those models um we didn't
36:18want people to uh go and have to find fine-tune those models but we are saying that hey you can still use any of those models that you have and then uh follow our pipeline we discussed in the paper so you can use that model still generate let's say four times 16 times more pixels >> and so talk a little bit about the generation process like what what's