4 ms·

Redream: Realtime Diffusion, Using Automatic1111 Stable Diffusion API

96 points by nateb2022 3y ago

AdmiralAsshat 3y ago
Isn't the name "Redream" already the name of a commercial Dreamcast emulator?
- lostgame 3y ago
  My first thought as soon as I saw the title. It is the name of a Dreamcast emulator, however; as far as I know it is mainly free and offers some sort of cheap upgrade for certain features. It's not open-source, that's for sure.
- phire 3y ago
  Commercial? Not really. It's closed source freeware with a donation option.
tomalaci 3y ago
[flagged]
- chorizoking 3y ago
  While the repos are racially provoking bordering towards outright racist, I’m not sure why that’s even relevant. If this stable diffusion code is good then the code stands for itself. What point are you even trying to make with this context?
- theodric 3y ago
  Irrelevant to topic and counter to guidelines https://news.ycombinator.com/newsguidelines.html https://news.ycombinator.com/newsguidelines.html
- boredemployee 3y ago
  Nice try.
charcircuit 3y ago
At 1 FPS this isn't quite real time enough to be useful for real time use cases, but I can see how a realtime stable diffusion would be a fun tool to use when producing for live events.
- 7373737373 3y ago
  Combined with video interpolation techniques it may work: https://www.youtube.com/watch?v=sFN9dzw0qH8 https://www.youtube.com/watch?v=sFN9dzw0qH8
asynchronous 3y ago
Corridor Digital (YouTube channel) recently produced a video using almost exclusively img2img stable diffusion to redraw scenes into a cartoon/anime style. I wonder if something like Redream could have sped up the production process rather than batch img2img.
- brucethemoose2 3y ago
  That whole pipeline is kind of a performance nightmare... I think a much better approach would be raw diffusers, at the very least least. I played with setting up a vapoursynth img2img pipeline myself (with VS doing the motion compensation to help with the temporal latent blending)
bottlepalm 3y ago
What's the latest on temporal coherence?
- brucethemoose2 3y ago
  IIRC there is a new controlnet model that is supposed to help. But simple pre and post processing with latent blending can get you pretty far.
mdrzn 3y ago
Love all these updates that are moving the open source scene forward, can't wait for a self-contained .exe install to try it.
kramerger 3y ago
Is there a cpu-only version of this? The base SD model run okay on CPU, and reasonably fast with intels MKL backend. But getting the whole auromatic1111 package to run on CPU seems to be a bit more complicated
pipes 3y ago
Thought this was going to be the Dreamcast emulator https://redream.io/ https://redream.io/

96 points by nateb2022 3y ago

AdmiralAsshat 3y ago

Isn't the name "Redream" already the name of a commercial Dreamcast emulator?

lostgame 3y ago
My first thought as soon as I saw the title. It is the name of a Dreamcast emulator, however; as far as I know it is mainly free and offers some sort of cheap upgrade for certain features. It's not open-source, that's for sure.
phire 3y ago
Commercial? Not really. It's closed source freeware with a donation option.

tomalaci 3y ago

[flagged]

chorizoking 3y ago
While the repos are racially provoking bordering towards outright racist, I’m not sure why that’s even relevant. If this stable diffusion code is good then the code stands for itself. What point are you even trying to make with this context?
theodric 3y ago
Irrelevant to topic and counter to guidelines https://news.ycombinator.com/newsguidelines.html https://news.ycombinator.com/newsguidelines.html
boredemployee 3y ago
Nice try.

charcircuit 3y ago

At 1 FPS this isn't quite real time enough to be useful for real time use cases, but I can see how a realtime stable diffusion would be a fun tool to use when producing for live events.

7373737373 3y ago
Combined with video interpolation techniques it may work: https://www.youtube.com/watch?v=sFN9dzw0qH8 https://www.youtube.com/watch?v=sFN9dzw0qH8

asynchronous 3y ago

Corridor Digital (YouTube channel) recently produced a video using almost exclusively img2img stable diffusion to redraw scenes into a cartoon/anime style. I wonder if something like Redream could have sped up the production process rather than batch img2img.

brucethemoose2 3y ago
That whole pipeline is kind of a performance nightmare... I think a much better approach would be raw diffusers, at the very least least. I played with setting up a vapoursynth img2img pipeline myself (with VS doing the motion compensation to help with the temporal latent blending)

bottlepalm 3y ago

What's the latest on temporal coherence?

brucethemoose2 3y ago
IIRC there is a new controlnet model that is supposed to help. But simple pre and post processing with latent blending can get you pretty far.

mdrzn 3y ago

Love all these updates that are moving the open source scene forward, can't wait for a self-contained .exe install to try it.

kramerger 3y ago

Is there a cpu-only version of this? The base SD model run okay on CPU, and reasonably fast with intels MKL backend. But getting the whole auromatic1111 package to run on CPU seems to be a bit more complicated

pipes 3y ago

Thought this was going to be the Dreamcast emulator https://redream.io/ https://redream.io/