I love it if still on 1000 generate per image.🙂
Playground is getting better and better and i'm loving it, great work and thank you.👍
tysm for adding counterfeit to sdxl! 🎉🙏😊
It's great isn't it? Hard to beat for it's style! I think you will also like Niji SE!
For now Copax Timeless is my absolute favourite. Versatile indeed.
Is there something wrong with Playground AI now? It doesn’t seem like it has these features anymore.
On canvas they won't show up when outpainting, in board mode you have to select Stable Diffusion XL as the model for these filters to show up (right hand side of the screen). @@Anonymous42316
Sheesh! I just took atleast 2 weeks of break from PGAI and we have so many new features. Happy prompting people.
@@playgroundai yes sir! A surprise indeed. And I am excited to try out this new filters.
These are really cool filters! Thank you for the awesome showcase of these new features!
They are and they will only get better as SDXL matures. Appreciate the support!
This video covers so much!😍
It was longer originally because I had a lot to say hahaha! I cut down the 2nd part otherwise it would've been a bit longer.
Thank you for the work and progress you make😺👍
Step by step, Playground becomes more and more interesting 🤩🥰😺
Thank you very much for your work, it is really wonderful the use and the community that grows day by day.
It would be a good element also to add an (adetailer) (face, body and hands improvement).
Thank you very much again!
Great as usual, but I am confused by the term "Filters"? are they actually trained custom models or Loras? Thanks 👍
Yeah I can understand that, technically they are models, except pixel art which is a lora. In the early days of playground there were no custom models, just text based filters that shaped the style of SD1.5. When custom models were introduced I believe at the time we just didn't want to confuse people and let the term stick. The average user doesn't know all the details of how the models are made and trained. They just want good results. 👍
After starting the program a few days ago I noticed that some of my filters are labeled SD1.5 assuming an upgrade. On a side note do you have a current video on the purpose of "board" in the format of Playground. It seems to do what canvas does but wondering how to use it in conjuction with canvas. On older videos the format is completely different. Thanx again for your effort and it seems to go the extra mile vs other videos.
Sd1.5 is an older model but still popular, the latest is SDXL. I've been thinking of doing a video explaining the differences between the two. As for board vs canvas, board is more for quick generations and many of our mobile users use it since we don't have an app yet. Canvas comes to play when you want to do more editing or image manipulation. Most people that work on desktops or laptops tend to favor canvas but many people use both in some way. Board was our first interface, Canvas came after. Hope that clears things up?
Stability: we're getting rid of military images
Playground: let me improve it for you
WOW! I have been waiting for this! I want to see Kandinsky! Russia has a great AI image gen slowly taking shape.
Great to see the excitement and anticipation! There isn't a lot of noise on Kadinsky but that can change really quickly :)
Thanks for the updates. Any info when can we get folders to sort our images? It looks like a big chunk in our profile page and getting performance issues.
I'd suggest to give feedback via the website, or the feedback section on Discord.
Niji SE & counterfeit look cool, i want to generation characters like anime card game style art 👍🏼
thank you! i've tried them out and there're some amazing filters, but for some reason the interface doesn't allow to generate 1024x1024 with them
Not sure what you mean? I’m literally creating images in those dimensions right now. Are you getting errors? Can you elaborate?
@@playgroundai i'm getting some bug, i guess. as soon as i choose one of the new filters (any of them) 1024x1024 option gets disabled. when i opt for any of the old filters, 1024x1024 becomes available again
You're most likely on board maybe? We have 2 interfaces, Canvas which you see on the video and board which is for quick generations or mobile users. On board you can select SDXL. Canvas is done by default.
@@playgroundai yep, i'm on board. i'm quite familiar with playgroundai (genuinely grateful for such an amazing tool!), and i've been using both board and canvas, but i find board to be much more convenient although i work from desktop. and you're right: there's no issue with dimentions on canvas, i get the bug only on board
Hhhmmm very strange, it would be helpful if you could provide a screen shot and a brief description of the issue to our support team that would be great! Support@playgroundai.com
especiales todos...
analog diffusion and polaroid filters have not been added :(
Unfortunately not available at the moment and I haven't heard if the original developers of those filters will be producing SDXL versions. I'm sure there will be similar ones made eventually for SDXL.
Very cool. I like the Kandinski model in Plazmapunk very much and the pixel art filter looks legit. 👍
Thanks for the new update! i would love to see a lego filter added to Playground next. keep up the good work!
That would be fun wouldn’t it? I’ve seen users prompt Lego but I get it, it’s not the same 👍🏼
Playground Ai went from 30 free upscales a day to only 20 free upscales a month LOL
@@playgroundai no it was 30 a day. About 2 days ago it changed unfortunately.
@@coldgin2023 Yepp. It was so sudden and in the middle of my work. I was shocked. Like 20 a month isn't gonna help much. 20 a day? I could live with. But a month? That hurts.
@@cruzinbosco yeah that is a world of difference....basically 900 freebies a month down to 20 lol
though i have had my frustations in playground its amazing how creative you can be and just getting better i will be here for the long time not a short run cause it just getting better thanks a bunch to the team honestly considering getting the year plan i highly recommend as someone who has adhd it can be alot to take in at first but it worth it much better than dall.e
We hope we can lessen those frustrations! 2024 will be amazing, sit tight! Appreciate the support.
Why am I getting bad fingers and unrefined faces all the time even when I use SDXL with those filters? Any specific parameters apart from negative prompt?
Kind of hard to help with no context or examples. It could be a number of things, your prompt, the filter, dimensions etc.
It's always the fingers for me even with adding negative prompts.The faces always come okay for me.
There's an almost excessive number of filters related to colors, styles, and so on. I believe the real challenge lies in maintaining poses and, of course, facial expressions and hand gestures.
As the tech matures those things will improve however there are tools like controlnet that help with that and facial expressions you can prompt for.
Some of these are useful but still tend towards cartoon type imagery. I hope it is okay to say this. Here is a question: Can you get stable diffusion XL to work well with image-to-image? The results are still much better using 1.5. PLEASE dont touch the "image strength" values, they work very well and are the key to Playgrounds power as a professional tool .
At the moment Playground "Stable Diffusion XL" it really does not work well with image-to-image. There is a softness and artificiality to the results. Even with the new samplers that came on today. So far none of them look remotely realistic (with the exception maybe of LMS Karras). To see this, take a picture of human face into Stable diffusion XL and use a image strength of 70%. Now try it with SD 1.5. The results are dramatically better, more useful and varied in 1.5
How to add watermark in playground AI
You have to create it separately as a transparent png file then bring it into canvas and place on top.
It's crazy how we get 1000 free image generations per day. SDXL 1.0 even! Many other websites give maybe 10 free image generations per day or less or even just a limited amount in total you can blow through in 15 minutes. Often without specialized models even. Just the base versions. NightCafe looks like shit unless you select SDXL and that's expesnive there. Playground AI can even create pretty good images with SD 1.5, let alone when selecting a specialized model. SDXL is already better in its base form than any of those I tried, let alone with the new specialized models.
@@playgroundai I also wanted to add my appreciation for your free tier allowing enough free generations that I can actually learn how to use the AI tool better. The other sites e.g. Mage, PromeAI, offer so few generations that I just gave up trying to learn. Once I learn how to generate art that I can earn an income with, I'll likely upgrade to a pro package in order to generate the art at higher resolutions since 1024 is obviously too low for printing, but for now I'm glad that you are offering a chance for people to learn and get addicted to this stuff.
My invite has expired on your discord....in the past I posted that this might become a problem and if their was a way around this because when I go to discord in general and do a search for playground ai I get a warning stating that this is not a playground and future searching will get me tossed from discord. I believe you said that you were going to check on that.
You don't need to keep using the invite. Just log into discord. discord.gg/playground-ai-1013195759178498068
How to create anime with playground AI?
Using any of the anime style filters will help like the ones in the video, Niji SE and Counterfeit, Rev Animated is also very popular, but you have to learn how to prompt for anime. Go the main playground site, playgroundai.com and search for anime. You can get inspiration from many of the users but make them your own.
The new filters are 🎉... but that new upgrade to pro bar and making free users wait 2min per generation is going to massively hamper overall growth.(people buying pro version)... cuz ALMOST ALL are not gonna wait around to get addicted... thus can not be pushed into pro.... and only will be pushed away foreva
不知道什么原因,我几个月前生成的图片,现在继续不改变一切条件,相同的种子,生成图片的风格,画质完全与种子不同.种子人物像真实的,美观的,新的生成图片像泥巴人,画质没有光泽,.在这之前,生成种子时,可以生成无限美观图片.
Sir you promised about comic creation🥲
Tell us what your favourite filter is! I'd love to hear from you!
Love the update👍. Very hard to choose just one, i like all of them but NijiSe, ProtoVision and Counterfeit V1 are my top 3.
ProtoVision!
@@RoguePhantom666 Great choices! Really hard to pick a fav.
@vVinchi Me too! 👊🙌