@antlers_anon ! Holy shit mate this is your best work.
Could you share a bit more details about the process? I am completely stunned at the quality of this work. Negative prompt, prompt weighting, process etc? Honestly this is the best AI art I think I've ever seen
@antlers_anon ! Holy shit mate this is your best work.
Could you share a bit more details about the process? I am completely stunned at the quality of this work. Negative prompt, prompt weighting, process etc? Honestly this is the best AI art I think I've ever seen
Thanks @chikubi, but I have no idea if anything I write would be helpful. Most of my process is just:
steal model ideas
- this one is modified berry,
steal parts of prompts (preferably from people not generating anime/porn)
- negative here is mostly something I found on reddit on a realistic painting-like image and cleaned up, - initial part of the positive is something I took from one anon who was generating greenskins, I just added my general tags and artist (obviously also changed "green" to "red" and "goblin" to "demon"), - "splash art", "complimentary colors" and "by Killian Eng" is from a post where someone was trying to imitate midjourney style with base SD.
generate a lot once I have something promising
- this applies to this prompt in particular. I've been testing it against different prompts and negatives and in total generated easily over 200 versions not counting inpainting.
add stuff I don't like to the negative
- in 90% of the images she had a second face on her stomach, ergo "face" (didn't help much), - in some images "pussy" turned into a cat, hence "cat", - everything was crazy red, not only her skin - "red", - I also added "skin" to have a bigger chance of getting some clothes.
I started collecting some tips for myself but then I noticed I disregard them most of the time, so they're not really useful. Best tip I can give is "steal shit from others and make it your own shit".
Thanks for the reply, but it can't be pure novelai. I've tried them out and completely different pictures come out that don't even come close to being the same. I also own a lot of models (1.5TB) and none of the hashs match.
Thanks for the reply, but it can't be pure novelai. I've tried them out and completely different pictures come out that don't even come close to being the same. I also own a lot of models (1.5TB) and none of the hashs match.
Ah., sorry. It must be a merged model then. Usually, unless the uploader tags them all, it's not easy to know at a glance what exactly the combo is.
Where can I download this model? > nai-aminefull-final-pruned_0.8-trinart2_step115000_0.2-Weighted_Sum-merged Your art is amazing!
- get trinart and novelai from the links in the OP of any of the diffusion threads on /g/ or other boards - merge them (automatic1111's webui has a nice tool for this)
- get trinart and novelai from the links in the OP of any of the diffusion threads on /g/ or other boards - merge them (automatic1111's webui has a nice tool for this)
I am sorry, I am new here and got brain damage. What does OP mean in the context of booru? You mean the model tags? I don't see any links there? The thing is I can't find any links for trinart or novelai, I thought they are only proprietary web page services? I also saw on your other piece "nai_0.95" I tried to google it but it just gives me machine parts, or is it NAI model with 0.95 weight (or whatever it's called) when merging and NAI is just novelai?
Also nice work but you know that, I just wanted to ask do you retouch renders from your prompts afterwards or do you just pick best ones and engineer them to perfection?
I am sorry, I am new here and got brain damage. What does OP mean in the context of booru? You mean the model tags? I don't see any links there? The thing is I can't find any links for trinart or novelai, I thought they are only proprietary web page services? I also saw on your other piece "nai_0.95" I tried to google it but it just gives me machine parts, or is it NAI model with 0.95 weight (or whatever it's called) when merging and NAI is just novelai?
Maybe I'm misunderstanding you but you may need to get started with the basics. Go here and follow the guide in section 1. Section 2 has NovelAI instructions.
I am sorry, I am new here and got brain damage. What does OP mean in the context of booru? You mean the model tags? I don't see any links there? The thing is I can't find any links for trinart or novelai, I thought they are only proprietary web page services? I also saw on your other piece "nai_0.95" I tried to google it but it just gives me machine parts, or is it NAI model with 0.95 weight (or whatever it's called) when merging and NAI is just novelai?
Also nice work but you know that, I just wanted to ask do you retouch renders from your prompts afterwards or do you just pick best ones and engineer them to perfection?
By OP I meant initial info in the threads on 4chan.
About the models, check the guide redjoe linked. It should explain everything.
I generally generate a few versions, pick the best one, retouch it a bit, do a bit of img2img, then SD upscale and then do some additional corrections (mostly colors, contrast and sharpness at this point).
Ah yes, yes, my bad I missed that /g/ part for some reason, I thought there is some "OP" in aibooru I'm missing. Now it all makes much more sense. I did just that and it turns out that NovelAI model is a leak, so that's why I couldn't find a repo for it I guess ;). That guide only had torrent magnet file, it's the same one from 4chan btw. I just assumed all models used with SD are open source and tried to find repo. Thank you both for your responses!
Thanks! Nice to hear you like it. What I did is basically an updated berrymix with r_34 replaced by trinart, here's a recipe: 1. novelai * 0.95 + sd-1.5 * 0.05 (weighted sum) 2. set method to "add difference" - model A: mix_from_step_1 - model B: f222 - model C: sd-1.5 - multiplier M: 1 3. mix_from_step_2 * 0.8 + trinart2 * 0.2 (weighted sum)
I also used float16 if you want to get the exact same hash.
Thanks! Nice to hear you like it. What I did is basically an updated berrymix with r_34 replaced by trinart, here's a recipe: 1. novelai * 0.95 + sd-1.5 * 0.05 (weighted sum) 2. set method to "add difference" - model A: mix_from_step_1 - model B: f222 - model C: sd-1.5 - multiplier M: 1 3. mix_from_step_2 * 0.8 + trinart2 * 0.2 (weighted sum)
I also used float16 if you want to get the exact same hash.
I seem to be getting a slightly different hash, despite going float16. I also get totally different unrelated results... I suspect something is up with my A1111 or something. Thank you for your help though anyway!~
I seem to be getting a slightly different hash, despite going float16. I also get totally different unrelated results... I suspect something is up with my A1111 or something. Thank you for your help though anyway!~
Weird, is your webui updated? Do you get a completely different style or mangled results? More realistic or flat textures? Or is it just different poses but same style?