Easy Stable Diffusion XL in your device, offline • Hacker News Board

Re: Easy Stable Diffusion XL in your device, offline

#161

Post by starwin1159 » Mon, Dec 04, 2023, 6:59 AM UTC

Any prompt keywords website recommended?

Re: Easy Stable Diffusion XL in your device, offline

#162

Post by cchance » Mon, Dec 04, 2023, 3:42 PM UTC

Haven't gotten to test it but given i use CoreML on Comfy, i wonder if we'll see more optimizations and performance work on the back end of these platforms as more useful frontends come out. the 1-4it/s on a 512 image is just sad, and the 2-3s/it on 1024 is just sad in this modern day, hell the ANE can't even run SD 1024x1024 images on a Macbook Pro M3 :S

Re: Easy Stable Diffusion XL in your device, offline

#163

Post by cchance » Mon, Dec 04, 2023, 3:43 PM UTC

Vicinity9635 wrote: post #96

wayfinder wrote: post #93

Earlier quoted context omitted.

If you want an absolute beast, especially for this stuff, you probably want Intel + Nvidia. Apple Silicon is a beast in power efficiency but a top of the line M3 does not come close to the top of the line Intel + Nvidia combo.

Well this would just be the excuse. I'm typing this on a Ryzen 5950X w/32 GB of RAM and a 4090. So I guess I already have the beast?

Get M1 macbook, throw your ryzen with 4090 in a closet and use it as a remote API endpoint for comfyui lol

Re: Easy Stable Diffusion XL in your device, offline

#164

Post by drcongo » Mon, Dec 04, 2023, 4:36 PM UTC

liuliu wrote: post #24

rgbrgb wrote: post #14

Just installed, this is very cool. Local AI is the future I want (and what I'm working on too). A few notes using it... Pros: - seems pretty self contained - built in model installer works really well and helps you download anything from CivitAI (I installed https://civitai.com/models/183354/sdxl-ms-paint-portraits ) - image generation is high quality and stable - shows intermediate steps during generation Cons: - do…

You should check out Draw Things on macOS. It works well enough for SDXL on 8GiB macOS devices.

I've been generating stuff non-stop in Draw Things for a few days, it's very good. Agree with the comments elsewhere about the rather overwhelming UI, and I have only one feature request: let us input the number of images we want to generate - the 100 limit means I keep having to check if it's finished to restart it.

Re: Easy Stable Diffusion XL in your device, offline

#165

Post by cchance » Mon, Dec 04, 2023, 5:27 PM UTC

mthoms wrote: post #48

solarkraft wrote: post #44

I find it interesting that it requires 16GB of RAM on Windows but 32 on a Mac. Unfortunately that leaves me out ...

I think that's probably because RAM on Mac is shared with the GPU. On Windows, you need 16GB RAM plus 8GB on GPU.

This is the correct answer

Re: Easy Stable Diffusion XL in your device, offline

#166

Post by ProllyInfamous » Tue, Dec 05, 2023, 6:03 AM UTC

myself248 wrote: post #13

ProllyInfamous wrote: post #2

The 16GB (base model) M2 Pro Mini, despite its overall awesomeness (running DiffusionBee.app / etc)... does not meet Minimum System Requirements (Apple Silicon requires 32GB RAM). So now I have to contemplate shopping for a new mac TWICE in one year (never happened before).

When choosing a machine with non-expandable RAM, you went with the minimum configuration? That's a choice, I suppose, but the outcome wasn't exactly hard to foresee.

At the time, it seemed that ANY upgrade made the M2Pro mini cost-inaffective.

One of my coller Q12023 ChatGPT experiences was having it help me "reason through" which machine was most "upgrade-proof," dollar-for-dollar.

Now in Q42023, I would definitely had made the decision to purchase an M2 Studio (base model) instead — those additional upgrades (VS M2Pro mini config, sim.) were much more cost-effective. Overall, I'm extremely satisfied with my M2Pro base model.

Re: Easy Stable Diffusion XL in your device, offline

#167

Post by ProllyInfamous » Tue, Dec 05, 2023, 6:04 AM UTC

sophrocyne wrote: post #4

ProllyInfamous wrote: post #2

The 16GB (base model) M2 Pro Mini, despite its overall awesomeness (running DiffusionBee.app / etc)... does not meet Minimum System Requirements (Apple Silicon requires 32GB RAM). So now I have to contemplate shopping for a new mac TWICE in one year (never happened before).

https://github.com/invoke-ai/InvokeAI - runs on Mac silicon, can squeeze out SDXL images on a 16gb mac with SSD-1B or Turbo models.

Thank you for this link.

Re: Easy Stable Diffusion XL in your device, offline

#168

Post by airesearch69 » Tue, Dec 05, 2023, 7:26 PM UTC

i created an AI Modell (have like 20+ very similar face photos of my dream model) created with "Fooocus" Juggernaut XL SDXL)...

what is the best way to creat myself an "Modell" or "checkpoint" of my wished modell face?

i find tutorials but they seem complicated and outdated. thanks for any small tipp what things to research for actually method. Thanks

Re: Easy Stable Diffusion XL in your device, offline

#169

Post by airesearch69 » Tue, Dec 05, 2023, 7:27 PM UTC

brucethemoose2 wrote: post #23

I would highly recommend Fooocus to anyone who hasn't tried: https://github.com/lllyasviel/Fooocus There are a bajillion local SD pipelines, but this one is, by far , the one with the highest quality output out-of-the-box, with short prompts. Its remarkable. And thats because it integrates a bajillion SDXL augmentations that other UIs do not implement or enable by default. I've been using stable diffusion since 1.5 c…

i use the same. Any ideas where to find actually (not outdated) guides on how to create your own "modell" out of the most similiar picutures of my dream modell? want to use it further with the same face on it.

thanks for any tipp guys :)