Announcements — Graydient / PirateDiffusion / AnimeMaker

CHALLENGE! "WEIRD SPORTS" Win a month of PirateDiffusion PLUS with LoRaMaker.ai - simply render an unusual sports activity. It can be a real sport or something made up, have fun with it. The winner will be announced in February. Good luck! https://t.me/challengepiratediffusion

New Feature! Background Replace You can now use an uploaded image or rendered image as the background. It works similar to the /faceswap command: First save a ControlNet preset (or) Use a photo ID, then do the /bg /replace:name command on the target photo. It works best when both photos are the same resolution. Here's a step by step guide on how to use it: https://piratediffusion.com/free-background-remover-ai-stable-diffusion/ We've also updated our After Detailer (adetailer) guide with more examples and tips: https://piratediffusion.com/how-to-use-after-detailer-adetailer/ We've also added hundreds of new Checkpoints like <juggernaut8-xl> and cool new loras like <qwek-xl>. Find them here: https://models.piratediffusion.com/concepts

Two new features! After Detailer and Showprompt Compare https://piratediffusion.com/how-to-use-after-detailer-adetailer/ #1 Add /adetailer to your /render to improve faces and hands. Works best when used with negative inversions. #2 You can now easily view your recent renders side-by side. Reply to an image with /showprompt and follow the link

What should we work on next? Tell us by submitting ideas and voting on our road map https://trello.com/b/aXVd3Q8J/pirate-diffusion-short-term-roadmap

https://www.youtube.com/watch?v=vgkrbPdow1k PirateDiffusion Show | Ep 4 - ControlNet Tips & LCM Sampler Subscribe on Youtube In this episode, watch how fast we are able create our logo intro graphics with AI using Controlnet, an image-to-image command that takes the basic shape of a graphic and renders into anything you prompt. We also cover LCM, a futuristic new technology that speeds up image generation. You can batch create image and get 9 images back in under 10 seconds all day long, it's mad science: 00:00 - Intro feat. Cat Grooming Masters 01:00 - $200 AI prompt challenge - "The Feast" 05:15 - How we made the intro graphics: ControlNet 10:00 - Two ways to start a Control 14:00 - Edges mode 17:50 - Masks and debugging 18:20 - You can type /cg instead of /controlguidance 22:50 - New Feature - The LCM Sampler 27:00 - Depth mode 30:00 - Contours mode 32:00 - Showprompt tips 35:00 - Using /mask to create a skeleton 38:00 - Lesson recap 49:00 - Poses mode 41:00 - Skeleton mode 45:00 - Quirks and tips

Three new commands, three new tutorials! Learn how to get greater control of your composition and colors with VAE, Vass, and FreeU. Includes easy examples that you can copy and paste into your prompts https://piratediffusion.com/features/ Special thanks to Forsana for the research & FreeU write-up

THE GREAT COMMUNITY SURVEY RETURNS! WE WANT YOUR IDEAS! How are we doing? Give us a minute, and we'll give you better products! Please answer our quick survey Thanks in advance!

Updated Tutorial! Build your own LoRa Our LoRA Training tutorial has tons of new advice on good photo selection and how to write a effective captions. It's a cool skill to learn, and will also help you understand prompting by looking at the whole model building process in reverse! https://loramaker.ai/train-a-stable-diffusion-model-online-faq/

The holiday challenge is here! Ten random entries will get $20 each, and you can submit up to 10 images. A lucky person can win more than once. This one's easy - a gathering for some kind of feast. A feast of friends? Let's go easy on the cannibalism, but make it festive. Festivus, even. Guidelines: - You can submit up to TEN (10) images in any art style - SFW images are preferred, artistic nudity also OK - Humor/memes always OK - All AI tools accepted - All concepts/models encouraged Allowed tools / Restrictions: Entries must be AI-generated, the full prompt must be visible. You can forward it as two messages or one. Number Entries - TEN You can submit up to (10) If more than the allowed number entries by one person exists, ALL of the entries by that person will be disqualified. Please don't spam — curate your best entries, and don't go over. Where to submit: https://t.me/challengepiratediffusion —- Winners from the November challenge are being voted on now, we'll announce the winners next week. Render Ahoy!

Our support website has been updated! New videos, tutorials, commands and tricks await you Please let us know in VIP chat on what stuff you would like to see covered in more detail https://piratediffusion.com/faq

Telegram is back! Render-Ahoy!

It appears Telegram is having an outage at the moment. All of our systems are online and WebUI is responding quickly, so please use our websites while this blows over While we have your attention, our new product LoRaMaker launches this December! You'll be able to train your own Stable Diffusion mini models right in your browser https://loramaker.ai If you're a PLUS member and would like early access to the beta, please give @yanman a shout

New PirateDiffusion Episode! Almost a full hour on how to boost quality https://www.youtube.com/watch?v=XP0Bg_SAy6I 00:00 - We forgot to remind you to win $225 05:45 - Use /more to upscale and fix hands 06:34 - Add strength and size in a second step 17:24 - Switch art style while in mid composition 23:35 - Create a reusable recipe or use a preset 31:13 - Limit size in first pass to avoid duplicates 42:20 - When to use guidance and strength, and why 43:00 - New /vass feature for SDXL and cheat sheet 48:00 - Lower concept weights to avoid glitches Buy some merch! https://piratediffusion.com/merch

New Video, 300 more models, new cheat sheet! Watch it on Youtube Learn Spin, Translate, ShowPrompt and more prompt engineering tips. Our features page has new tutorials, and news on upcoming commands like /freeu * CONTEST! Our Fashion AI challenge ends soon, for $225 in prizes! Get that money! * https://graydientplatform.com/whats-new/

——- NOVEMBER CHALLENGE ——— =============================== Realistic AI Fashion Product =============================== $225 in Prizes! Judged by a real fashion creator in NYC $100 - Judge's Pick - Grand prize, Most fashionable $50 - Judge's Pick - Runner up $25 - Judge's Pick - Third place $50 - Random winner, participation prize Official Rules and where to submit: https://t.me/challengepiratediffusion Good luck!

Last call! Enter our Halloween Challenge: $50 cash prize and $50 random draw https://t.me/challengepiratediffusion (pictured: entry by AiUnlimited)

New! WebUI is now called Stable2go, and it has the much-requested dark mode. Here's an early look at the new colors: https://stable2go.ai/concepts We're also in the process of creating new tutorials and video guides. Our new features section splits Telegram and Stable2go videos into two separate areas. https://graydientplatform.com/whats-new/ More focused Telegram-only tutorials are on the way without trying to cover both at once. We hope you like the new look and would love your feedback

Reminder: There's a Halloween Render Challenge going on, $100 prize split for participation and judges choice, get your render in! Bring your peekaboo gimps! (Prompt by Eggs Benadryl) Enter here: https://t.me/challengepiratediffusion

The PirateDiffusion Show - Episode 1 Intros & Halloween Contest - 00:00 We now have an API - 04:00 The basics - making cats with chats - 04:30 Loading models to change the art style - 06:41 Our favorite LoRA of the moment - 08:00 Intro to LoRA weights - 10:00 Community Questions: Recipes, Habits - 15:00 ControlNet Basics - 24:30