Iconography and stained glass style are super fascinating to me. I'm in love with the 2D/2.5D style of Magic The Gathering artist _Seb McKinnon_ , I was astonished to find that PlayGround recognized his style immediately.
Doodles of all kinds. Everything I’ve tried so far does not produce a good result. Cute tiny doodles, minimalist with fine outlines and minimal muted/pastel colors, with a charming vibe. Also basic outlined vector-style neat clean doodles. Also scrappy/scribbly/ruffian doodles with rough outlines in a dry brush look with fills looking chalky. Also whimsical doodles with a stick-figure vibe. And patterned doodles that can be filled in with different colors very neat clean precise.
I did not even generate 50 images of the playground. But now I am not allowed to generate. When I press the generate button this message keeps coming up. I tried again every day for two or three days. But this message keeps showing. The message is, "Your queue position is based on the number of images created and demand is high right now. Please try again later or upgrade"
I think the period of free playground is coming to an end. XL is constantly in queue. I could generate images on 1.5, but since they are removing it on March, I dont think unpaid version will be usable.
Just wanted to say I love all these videos, one video I would really like to see, is one that focuses on how to choose the right filter and model. as one wouldn’t want to go through each filter for both sdxl and PGv2 until you find the right style for what you’re doing
I am in awe of how generously informative the videos are. This is exactly what I have been searching for. What recommendations can you share for avoiding color, grayscale and shading when prompting for a coloring book style image of an Afro-Latinx or an African-American person?
That would be huge! I Think once we get Playground V2.1 out and V3, this could be an area of focus as it's not widely done at the moment. Only a couple other companies I know of that does it well.
Do you have any tips for controlling color better. ControlNet is awesome, but I find color can get out of hand and adjective bleed becomes a huge problem.
Generally controlling specific areas through prompting can only get you so far but it does help to be specific in those kinds of details. e.g black leather jackets, blue jeans etc. But when it comes to inpainting or masking, it can be tricky and tedious but there are some techniques that can work depending on the context. I'll share something soon on this.
@@playgroundai Thank you for your reply. Color is typically very good, but can be challenging at times. As always, I look forward to anything you can share on this or any other topic. Thank you for taking the time to make these videos and answer questions.
Hi, I wanted to ask a quick question: Do weight work in Playground AI? For example, if I have (car:1.5) in my prompt, will it put more emphasis on the work "car" ?
está muy bueno, hice cuatro imágenes, ahora tengo que esperar , ya que lo utilizo gratis, y el servidor está ocupado. Unas en Stable Difussion XL (las mejores y filtro Starlight XL)y las otras dos en DPM++2Karras con el filtro Juggernault XL una salió bien pero con un fondo como de fotografía un poco amarillento, y la otra el rostro era realista jaja, pero bien, hay que probar, gracias por las indicaciones, saludos
No matter what I do I can't get these kinds of results. Everything with juggernaut comes out looking realistic. All other images come out with greyshading. If I use the playground model it's more detailed but I can't get clean lineart it has lots of greyshading. More research needed for me. I need to create a preset so I can create some coloring book pages
I got good results using Stable Diffusion XL with Juggernaut & Eular a, with this type of prompt: Very simple, very basic, vector thin black outline, smooth, … minimalist. Also with Real Cartoon XL & DPM++ 2M SDE Karras, with this type of prompt: … hand-drawn illustration style, linear art, black, isolated on white background. Also I just tried this with the last mentioned filter and it worked well to get a simple outline: “Extremely simplified reduced black vector outline of a kawaii flower, for toddler’s coloring book”. So just play around to see what works, and don’t forget to put things like this in the Exclude From Image for negative prompts: detailed, complex, intricate. Good luck with your experiments.
@@justjules5831 I re-did some images and used your promt as a basis as wel las the settings and the images came out alot better. Gonna try that later now in foocus and Comfy UI. Thanks
"Your queue position is based on the number of imaged created and the demand is high right now. Please try again later or upgrade to pro." Really? There are lot of people creating images at 12:17 am EST and at 10 am as well? Strange....
How to generate a 3d icon with different texture like glass, metal. It should be look like a build in blender and cinema4d, I am trying but didn't get desired result