There’s probably no reason to be running two Pis. A single one could easily do both.
I just recently got myself a Pi 4 8GB for €86 on Amazon, a barely used 10TB IronWolf NAS HDD for €120 on eBay, and an HDD enclosure for €20 that powers the HDD and plugs into the Pi via USB3. Currently it’s running Plex Media Server, qBittorrent, Wireguard, AdGuard Home, Nextcloud, Caddy, and a few monitoring containers. It’s handling all of that easily. I’m using it headless and just stream content to my TV. Currently I have no redundancy for my drive, but I’ll get around to that eventually.
You can of course keep using PiHole for your setup. I recommend you check out !selfhosted@lemmy.world and r/selfhosted.
You don’t even need to buy a microsd card anymore because they now ship with the firmware that can boot from USB.
My purpose is fulfilled.
Like Gooey Louie?
Thanks for sharing the detailed workflow
Gotta love that horrific blobby mass with fingers jutting out and twisting at all angles.
Thanks, it was a little difficult to get it because Midjourney kept blocking my prompts because of words like ‘disfigured’. Kind of ridiculous. It was also hard to get a blonde woman in a bed without it getting blocked.
You may be able to find an artist with a similar style here https://www.midlibrary.io/categories/illustrators
Goldilocks’ Rude Awakening
Midjourney provided the proper composition and SD (RealCartoon3D) jazzed it up.
The Truly Ugly Duckling
All Midjourney for this - in the style of Arthur Adams.
The Longings of Rapunzel
For this one, SDXL provided the composition and it was greatly improved with the Fruit Explosion model.
Deep Space Wonders
That looks pretty cool, with more features than I thought it would have. One thing I noticed is that it doesn’t seem to support embeddings and while it has a lot of models, it’s missing some really popular ones such as Rev Animated. Could be a good start for getting ones feet wet.
For the first image I used Arthemy Comics and for the second one I used BrickAndMortarMix. Of course I also used a whole mess of loras and embeddings.
I’m using Stable Diffusion. While simple generators, such as Midjourney can produce pretty good results, your control over it is extremely limited.
For this, you have several options, but only a two ways actually give you 100% control. The first is to run it locally, but this requires a pretty good GPU. The second way is to use a GPU cloud service, such as runpod.io or vast.ai. I’ve used both, but recommend the former for beginners as it is more user friendly. You can rent an RTX 3090 for 0.44/hr, for example. I don’t recommend using the base model for Stable Diffusion because the results are pretty meh. Civitai.com is where you can find a basically unlimited amount of resources including models.
There are a few user interfaces for it, with the most used being Automatic1111. If you’re installing it locally to use with your own GPU, check this video https://www.youtube.com/watch?v=kqXpAKVQDNU. If you want to try runpod, check https://www.youtube.com/watch?v=9Tf5WBXyuOI and if you want an EXTREMELY in-depth video https://www.youtube.com/watch?v=QN1vdGhjcRc.
It can be overwhelming at first, so I recommend checking out three of my favorite channels to get you started with generating images:
The Terminator Mk.II and Lunar Base are ReV Animated v1.2.2 and the Terminator Mk.1 is Arthemy Comics v4.
I’m a bit late, but here are my submissions, all made with Stable Diffusion.
Dieselpunk Terminator Mk.I
Dieselpunk Terminator Mk.II
Dieselpunk Lunar Base
Thanks!
Oh, that’s interesting. I probably went through over hundred generations of the face so I may have become desensitized to small oddities.
Are you referring to the glowing eye? I purposely inpainted it that way 😅
There is some resemblance isn’t there