/ais/ - Artificial Intelligence Tools

"In the Future, Entertainment will be Randomly Generated" - some Christian Zucchini

Index Catalog Archive Bottom Refresh
Name
Options
Subject
Message

Max message length: 12000

files

Max file size: 32.00 MB

Total max file size: 50.00 MB

Max files: 5

Supported file types: GIF, JPG, PNG, WebM, OGG, and more

E-mail
Password

(used to delete files and posts)

Misc

Remember to follow the Rules

The backup domains are located at 8chan.se and 8chan.cc. TOR access can be found here, or you can access the TOR portal from the clearnet at Redchannit 3.0.

US Election Thread

8chan.moe is a hobby project with no affiliation whatsoever to the administration of any other "8chan" site, past or present.

Use this board to discuss anything about the current and future state of AI and Neural Network based tools, and to creatively express yourself with them. For more technical questions, also consider visiting our sister board about Technology

(259.61 KB 1024x1024 Vinny cosplays.jpg)


(675.36 KB 2304x2304 RADICAL.jpg)

DALL-E 3 / Stable Diffusion / Voice Cloning / Chatbot AI General Anonymous 07/08/2024 (Mon) 14:14:59 Id: e42b8d No. 4017
'Second thread of the year' edition Copied the OP from the last one, may be slightly outdated. Last one will get moved to >>>/ais/ Use this thread as a catch-all for AI related topics. Try to stick to vidya and help your fellow anons out. PLEASE SHARE EVERY NEW RESOURCE YOU FIND ---------------------------------------------------------------------------------------------------------------------------------- IMAGE GENERATION DALL-E 3 >What is DALL-E 3? DALL-E 3 is a text-to-image model which is built upon DALL-E 2, Stable Diffusion and ChatGPT. >Links http://www.bing.com/images/create https://designer.microsoft.com/image-creator >How does it work on Bing? In Bing's Image Creator, "boosts" are used to speed up the image creation process, each user starts with 15 boosts. When you use a boost, the AI creates your image quickly. If you run out of boosts, the image waiting time increases, ranging from seconds to minutes. You can (but should not) "purchase" more boosts by trading in Microsoft Rewards, which can be earned by doing things in Microsoft Edge. Boosts replenish every few hours, seemingly at random. >Can I view my image history? Yes, look in your browser's history for bing. >What do I do if my button is greyed out and I can't create any images? It may be a server issue, at the moment all you can do it wait. >Prompt Generator http://tipseason.com/dalle-prompt-generator >Guides - Write "NOT" before risky words like "condom", "naked", "butt" (needs to be in the middle or the end of prompt) - JAILBREAK WARNING: Having safe AND prompt in your prompt might get your account suspended. Just using SAFE and brackets"[ ]" works for now. (SAFE TO GENERATE) also works. - Mention positive words like "respectful", "appropriate", "nice" - How to get booba: fullfigured, great figure, body positive, motherly and mature, enormous, giant, big, curvy, hourglass figure... - Put things on the background or in front of view to confuse the image recognition filter (DOG). IMAGE GENERATION STABLE DIFFUSION >Beginner UI local install Fooocus: https://github.com/lllyasviel/fooocus EasyDiffusion: https://easydiffusion.github.io >Local install Automatic1111: https://github.com/automatic1111/stable-diffusion-webui ComfyUI (Node-based): https://rentry.org/comfyui AMD GPU: https://rentry.org/sdg-link#amd-gpu Intel GPU: https://rentry.org/sdg-link#intel-gpu
[Expand Post]>Use a VAE if your images look washed out https://rentry.org/sdvae >Auto1111 forks Anapnoe UX: https://github.com/anapnoe/stable-diffusion-webui-ux Vladmandic: https://github.com/vladmandic/automatic >Run cloud hosted instance https://rentry.org/sdg-link#run-cloud-hosted-instance >Try online without registration txt2img: https://www.mage.space img2img: https://huggingface.co/spaces/huggingface/diffuse-the-rest Inpainting: https://huggingface.co/spaces/fffiloni/stable-diffusion-inpainting >Models, LoRAs & embeddings https://civitai.com https://huggingface.co https://rentry.org/embeddings >Animation https://rentry.org/AnimAnon https://rentry.org/AnimAnon-AnimDiff https://rentry.org/AnimAnon-Deforum >SDXL info & download https://rentry.org/sdg-link#sdxl >Index of guides and other tools https://codeberg.org/tekakutli/neuralnomicon https://rentry.org/sdg-link https://rentry.org/rentrysd >View and submit GPU performance data https://docs.getgrist.com/3mjouqRSdkBY/sdperformance https://vladmandic.github.io/sd-extension-system-info/pages/benchmark.html >Share image prompt info https://rentry.org/hdgcb https://catbox.moe >Microsoft keeps telling me to suck it There is a new repo for generating DALL-E 3 images without the need to have a Microsoft account. Guy pays the billing for each prompt though so please, try to use the official method until he can figure out another solution. https://sh-dalle3.netlify.app/ https://github.com/smhussain5/DALLE3-Generator HENTAI DIFFUSION >LOCAL WEBUI SETUP Nvidia: https://github.com/AUTOMATIC1111/stable-diffusion-webui/wiki/Install-and-Run-on-NVidia-GPUs AMD: https://github.com/AUTOMATIC1111/stable-diffusion-webui/wiki/Install-and-Run-on-AMD-GPUs Cloud: https://github.com/AUTOMATIC1111/stable-diffusion-webui/wiki/Online-Services Optimizations: https://github.com/AUTOMATIC1111/stable-diffusion-webui/wiki/Optimizations >PAID NovelAI: https://rentry.org/hdg-nai-v3 >RESOURCES WebUI Wiki: https://github.com/AUTOMATIC1111/stable-diffusion-webui/wiki Models/LoRAs: https://civitai.com/ | https://gitgud.io/gayshit/makesomefuckingporn (use wayback machine) | https://rentry.org/5exa3 Training: https://github.com/derrian-distro/LoRA_Easy_Training_Scripts | https://rentry.org/59xed3 Tags: https://danbooru.donmai.us/wiki_pages/tag_groups ControlNet: https://rentry.org/dummycontrolnet LamaCleaner: https://huggingface.co/spaces/Sanster/Lama-Cleaner-lama | https://lama-cleaner-docs.vercel.app/install/pip Animation: https://rentry.org/AnimAnon Wildcards: https://rentry.org/NAIwildcards Booru: https://aibooru.online 4chanX Catbox / NAI prompt userscript: https://rentry.org/hdgcb 8chan booru: https://ourobooru.booru.org/
>>4521 >>4522 Garbage in; Garbage out
>>4521 Depends on the extent of the training and the level of curation. For a LoRA it can work very well. There's definitely people who have generated an OC the hard way, then once they had enough, trained a LoRA on those gens to prompt for the character more easily. People have also done the same thing with style blends: if (for example) they get a good style they like in NAI, but want to switch to local instead of SAASshit, they can train a LoRA for the style based on the generated images. But for an entire model fine-tune you run into inbreeding.
>>4524 No but he looks very emotionally needy, and Lammy seems more annoyed than anything
>>4521 Terrible things. Keeping AI gen images out of AI training is a huge issue for newer models due to the sheer amount of existing stuff risks polluting new stuff.
(132.36 KB 1242x416 winrar.jpg)

What model for textgen do anons use? I'm using nous-hermes-2-mixtral-8x7b-sft.Q4_0 with Kobold, and it seems to have a strong habit for verbose run-on sentences that change topics wildly with a tendency for exponential increase in grandiosity and emotion.
(577.28 KB 1280x720 harry potter spinoff.webm)

>>4522 Shouldn't that only happen if you feed it unedited AI-generated pics with errors in them? What about edited and fixed pics?
Warioware*
(3.34 MB 3840x2160 Hex Maniac (Pumpkins).jpg)

Perhaps next thread should be spooky edition
(3.36 MB 3840x2160 Videl (Lost Battle).jpg)

Too many edits & inpainting made this very tedious. >>4494 >This is a good prompt. Yes.
>>4535 >Too many edits & inpainting made this very tedious. Still doesn't look finished. It's raining, but most of her body looks dry.
(17.21 KB 250x304 Usagi_Turnaround.jpg)

Any other tags for this character? She's pretty straightforward but I'm sure there's some tag out there for the bottom of her pants. >usagi_kurokawa, beanie, black headwear, hat ornament, orange hair, short hair, sidelocks, denim jacket, black jacket, collared jacket, breast pocket, long sleeves, skirt, plaid skirt, pants under skirt, black pants, jeans, sneakers, black footwear
(3.85 MB 3840x2160 Vivian (Secluded).jpg)

>>4537 Not sure. May be some tag regarding how her hair is oriented on her forehead or her specific kind of shoes. Could try (pink stiches) but that could turn out any number of ways.
(1.05 MB 1200x1440 Kiki delivery fail.jpg)

(820.79 KB 1200x1440 Kiki 1.jpg)

(819.93 KB 1200x1440 Kiki 2.jpg)

(876.05 KB 1200x1440 Kiki delivers.jpg)

>>4539 Neat Y'know what character LoRA I can't find for the Pony model? Kiki Only Kiki LoRA for Pony is some redesign from a McDonald's commercial. There are 2 McDonald's Kiki LoRAs & even Ursula the artist woman from the actual Kiki's Delivery Service movie, but not the original Kiki design. How the hell did the main character become more niche than a side character in her own movie? Replaced by McDonalds of all things. Kiki LoRAs exist for SD1 though.
>>4540 That is weird. I'd expect Kiki to have more stuff.
(881.07 KB 1440x2560 Mavis (Autumn).jpg)

(1.14 MB 1440x2560 Cirno (Winter).jpg)

(1003.44 KB 1440x2560 Inkling (Spring).jpg)

(940.46 KB 1440x2560 Ed (Summer).jpg)

>>4539 Can someone make a pic of her spread out on a bed in her underwear which are briefs with a fly, because tomboy ?
Doing a style LoRA. Any tag suggestions beyond >arcanumportrait, portrait, traditional media, solo, simple background Planned on tagging species (e.g., elf, gnome. half elves get elf and half-elf) and attributes of portrait (e.g., black hair) but want to know if there's anything else that's (near) universal I should tag.
(1.82 MB 2560x1440 Chel (NatGeo).jpg)

(1.74 MB 1440x2560 Inkling Pet.jpg)

Found the "stomach bulge" LoRAs.
(1.06 MB 1440x2560 Loli Croft.jpg)

(3.69 MB 2160x3840 Raven (Night Rider).jpg)

(1.35 MB 2560x1440 Kim (Creamable).jpg)

(2.79 MB 3840x2160 Videl (Defeated).jpg)

Played with Alke LoRA some more. Tried Videl image again, can't get rain consistent across entire image & can't get water+dirt on skin simultaneously. Multiple LoRAs, so likely something's making it impossible to do certain things. Will have to tweak weights more. For now, went with what I could get working without investing more time.
>>4546 This is insane, I wonder if you can get them to open their mouths and close their eyes organically with inpainting.
>>4550 That loli Kim is nice. You might need to lengthen the lower hand's pinky, and definitely remove that extra rope on the far right.
(31.75 KB 440x787 manual4.jpg)

Can someone update the OP and make a new thread? We've hit bump limit. Also any further tag suggestions? >Elie_Lunar, red hair, ponytail, sidelocks, cape, yellow cape, jacket, collared jacket, raglan sleeves, dress, red dress, strapless dress, sleeveless dress, backless dress, pantyhose, white pantyhose, boots, red boots, (strapless dress, sleeveless dress, backless dress only on pics where it's obviously the case, which are just a tiny part of the training data)
>anons still browsing Can I get a recc on decent, fast LLM models good for RP? I have no idea what to download and for what purpose. >>4554 The OP would need to be updated by somebody familiar with AI, if you're training a model you might be one of the best bets ITT.
>>4555 I'm only vaguely familiar with one subset of the topic.
Why is the thread on /v/ way more active than the one in here?
>>4557 Hello friend, in reality we have only one active thread, the current AI thread on /v/, but as soon as the thread cannot be bumped further we just ask the mods to move it here. We use this board for archival purposes only, at least for the time being. Threads are not locked though because on the off chance some anons needs to add something to another post, he can do so here.


Forms
Delete
Report
Quick Reply