Photoshoot in Colter +LoRA [Album]

3 years ago by Cavendish to c/aigen

Today, we're headed to the frozen north. Dispite the snow on the ground, the sun is out and the light is perfect for a brisk shoot at the weather-worn cabins of Colter.

Two months ago, I fell into the trap that is Stable Diffusion. Today, I released my first trained model based on the snowbound town of Colter from Red Dead Redemption 2. For anyone interested in SD image generation, you can grab a copy at CivitAI. https://civitai.com/models/137327. I'd appreciate you taking a look, and giving it a like or a rating if you're so inclined. The LoRA model is stylistically versatile, and there's a bunch of SFW examples I made of its range.

As always, images link to full-size PNGs that contain prompt metadata.

DoucheBagMcSwag 9 points 3 years ago

I…have got to learn more about this AI shit….

path: 0 2971124, hotness: undefined, score: 9, children: 3

Cavendish 5 points 3 years ago

path: 0 2971124 2972455, hotness: undefined, score: 5, children: 2

DoucheBagMcSwag 2 points 3 years ago

Too time consuming?

path: 0 2971124 2972455 2976849, hotness: undefined, score: 2, children: 1

amotio 2 points 3 years ago

Yes

path: 0 2971124 2972455 2976849 2978088, hotness: undefined, score: 2, children: 0

RandomUser88 7 points 3 years ago

Thanks for your contributions to the community!

I have questions if you don't mind.

In really trying to get into LoRA training in general and there's a lot of things I can't intuitively work out or find solid answers to.

For example, with this, what is your "class"? And did you use regularization images? (I want to make a habit of using them). If you did, what did you use for them? Like places that aren't this? Like deserts and forests, etc?

Would you consider elaborating on batch size, repeats, epochs, etc, too?

Thanks again!

path: 0 3011768, hotness: undefined, score: 7, children: 3

Cavendish 5 points 3 years ago

path: 0 3011768 3013236, hotness: undefined, score: 5, children: 2

RandomUser88 2 points 3 years ago

Oof. Dude. You're not wrong about what is and isn't available online. But it's okay. New frontier or whatever. Haha.

I've been mulling over the regularization image thing, so I created a reddit post asking about it, but I basically asked, "are these images supposed to represent what the model thinks 'this' thing is, and in that case, regularization images would serve the role of being 'this, but not this'" or is it more like, "these fill in the gaps when the LoRA is lacking?"

I suspect it's more like the first. That said, it might actually make sense to include all the defective and diverse images for the purpose of basically instructing the LoRA/model to be like, "I know you think I'm asking for 'this,' but in reality, that's not what I want."

If that's the case, it might make sense to ENSURE your regularization images are way off base and messed up or whatever. Or at least anything in the class that you know you def don't want.

I don't have confirmation of any of this. I'm VERY new here (like ran my first LoRA training yesterday).

I like the idea of your batch size.

Ah. The captioning is something I REALLY need to think about. I'm guessing the cabin caption idea you used, basically you lost flexibility but gained accuracy by going that approach? I wonder if you could tag it 'cabin, church' and retain some of both?

The steps, to me, sound very high, but I can't say, for sure. Ahaha. Because for people, I've heard 1500 to 3000.

I'll be sure to come back and share findings once I have more. I think to really "do this right" you HAVE to train some of your own shit, but to do it well, as you've quickly realized, you've got to understand the methodology/philosophy of how it's done.

path: 0 3011768 3013236 3014105, hotness: undefined, score: 2, children: 1

RandomUser88 3 points 3 years ago

Well. Maybe scratch some of what I said above. As with many things, the answer is simply more complicated than that.

I found this video fairly useful in helping understand the process. I hope it helps.

https://youtube.com/watch?v=EehRcPo1M-Q

path: 0 3011768 3013236 3014105 3016117, hotness: undefined, score: 3, children: 0

7112 5 points 3 years ago

Thanks for always sharing the process. These are really good. The backgrounds look great. Always impressed you pull these from games.

path: 0 2965571, hotness: undefined, score: 5, children: 1

Cavendish 3 points 3 years ago

path: 0 2965571 2970834, hotness: undefined, score: 3, children: 0

lobster_fowl 4 points 3 years ago

First, another great album, thank you!

One thing I always find interesting is to look at the clothes generated, and they're usually kinda bizarre, but I the jacket/jumper thing from #6 would actually look pretty cool in real life

path: 0 2970318, hotness: undefined, score: 4, children: 1

Cavendish 3 points 3 years ago

path: 0 2970318 2970788, hotness: undefined, score: 3, children: 0

uncle_dewey 4 points 3 years ago

This is really amazing work!

path: 0 2975232, hotness: undefined, score: 4, children: 0

Kodemystic 2 points 3 years ago

Hey these look great and that 1st one is just... wow . What model do you use?

path: 0 3014949, hotness: undefined, score: 2, children: 1

Cavendish 1 point 3 years ago

path: 0 3014949 3030584, hotness: undefined, score: 1, children: 0

aigen

@lemmynsfw.com

login for more options

Welcome to the LemmyNSFW home of AI generated images and (for the wizards that can make them) animations of both the anime and realistic nature.

If you're here to share resources & ask questions, please use the stickied discussion thread for that.

Remember to mark your posts as NSFW.

Rules

Just don't be a dick, really
No depictions of celebrities and other actual living people.
- Without explicit and verifiable consent
Consider sharing your prompts. Not required, but it's nice, y'know
- Lemmy will strip out the prompt data from images. Consider either uploading somewhere else (Like catbox.moe) or manually pasting the data into the body before posting
No disturbingly inaccurate "body horror" generations
No advertising of any kind.

And of course, remember to follow LemmyNSFW's rules (duh)

Notice

AIs don't know the law, and can't judge their outputs based on how... legal... they may be. Exercise caution and generate responsibly. If you see anything suspicious, report.

Also Check Out

go to feed...

aigen

@lemmynsfw.com

login for more options

Welcome to the LemmyNSFW home of AI generated images and (for the wizards that can make them) animations of both the anime and realistic nature.

If you're here to share resources & ask questions, please use the stickied discussion thread for that.

Remember to mark your posts as NSFW.

Rules

Just don't be a dick, really
No depictions of celebrities and other actual living people.
- Without explicit and verifiable consent
Consider sharing your prompts. Not required, but it's nice, y'know
- Lemmy will strip out the prompt data from images. Consider either uploading somewhere else (Like catbox.moe) or manually pasting the data into the body before posting
No disturbingly inaccurate "body horror" generations
No advertising of any kind.

And of course, remember to follow LemmyNSFW's rules (duh)

Notice

AIs don't know the law, and can't judge their outputs based on how... legal... they may be. Exercise caution and generate responsibly. If you see anything suspicious, report.

Also Check Out

go to feed...