First, I would like to thank the Playground family for this unique site Which I expect to become the most important and powerful tool for AI Second there are some issue need to be modified, the most important of which is browsing the private profile and members’ pages. Browsing is very slow and often stops. I suggest dividing the profile into several pages, like Flickr, so that each page contains 100 photos max. This modification is absolutely necessary to improve the user's interaction with the site, thank you
I'm glad to hear that Richard. There is so much more to cover and what we can do with ControlNet. It's important to understand the basics before you can get creative with it. More to come soon!
This video is worth millions! Because I don't have a premium account, I waste a lot of time generating dozens of images that in the end won't do! And on my own, I've tried every way to use this function to generate different poses, maintaining consistency, and I've never been able to do it! I'll try it now using your tips! Thank you
Thanks so much for these great tutorials! I had not experimented with the ControlNet image controls but using edge control is already generating images that are amazingly on target with my goals. The image pose seems largely ignored in the prompt, but edge controls works beautifully (and I just started experimenting so much more to go but already blown away with results).
Thank you for this. What do you do with Edge's threshold? From 1-255. I had a prompt which I wanted to use a Pitbull to replace the cat riding the bike in my image prompt. Can this be done successfully? With what I did, I seemed to get lots of deformities. I used those #'s you used- .6 for Edge & .4 for Depth. Guess practice makes perfect.
Tnx for clear explanation. By my opinion, it's lack of possibility to communication in Playground community. I would be like to have possibility to communicate not only in share on canvas.
Yes, combining them will usually yield better details and more control, it's not always necessary and dependent on the reference image. So much to show you all still! 🙌
I really like playground AI and the features you guys add. But I wish if you could add more filters/checkpoints like adding some more anime style filters would be great. hope you guys consider this suggestion.
Absolutely! I'm always looking for new ones to suggest to the team, as long as there are no licensing issues. Do you have any favorite's that you like?
Very good tutorial! As always. ;) Thank you for all of these. But I have a question regarding the additional control at "Edge". There's a second control slider besides "Control Weight" called "Edge Treshold". Although I have an idea what this is doing, I would like to ask, what is this thing doing exactly? And what are good tresholds to have to get certain results?
Thank you for the great tutorials as always. I'm learning how to use Playgroundai to color my coloring book pages. I've seen Stable Diffusion users do it with 1 page of manga, but I don't know if it's possible to use Playgroundai and Controlnet to do it? I also want to do the reverse: turning pages of colored pictures into coloring book pages. If it's possible, I hope you can guide me.
Hi, I don't have "control traits" on board anymore but I've finally found time to experiment with this. Where did the settings for it go? (I have a pro account)
You can only use controlnet with stable diffusion 1.5 at the moment, switch the model on the top right to stable diffusion 1.5. ControlNet for SDXL will be coming soon.
@playgroundai thanks so much. What happened to the playground ai model in the top right? Has it been discontinued? Is there any comms about changes like this I can sign up to? Cheers
Yes there are methods but I will say, generally speaking the nature of stable diffusion is to create random images so achieving 100% consistency if almost impossible. With that being said there are work arounds. Vids to come soon on this! Thank you for the support!
not without training which we don't have yet however if you use a fictitious name or even 2 together in your prompt, you will get a consistent face. I'll be doing a video on this soon.
@@playgroundai playground v1 or standard stable diffusion1.5 but for free version has only playground v2 and standard stable diffusion XL , so the free version cannot use because the model isn't avaliable for free user , am I understand it correctly?
@@playgroundai I also cannot adjust the control as pose 0.6 edge 0.4 dept 0.4 , it’s all run together , please let the team check it thank you very much for the team support all issue
The graphic was for demo purposes only, if you watched this part 2:19 I explain how that same reference "pose" shows up in the thumbnail of controlnet.
![ControlNet - Openpose [TensorArt]](http://i.ytimg.com/vi/vGIKA_K3RzQ/mqdefault.jpg)
![ControlNet - Openpose [TensorArt]](/img/tr.png)
Awesome!
Thanks!
Very helpful. Thank you!
Glad it was helpful! 👍
Playground and your teaching is wonderful 🎉
You are so kind thank you!
First, I would like to thank the Playground family for this unique site Which I expect to become the most important and powerful tool for AI
Second there are some issue need to be modified, the most important of which is browsing the private profile and members’ pages. Browsing is very slow and often stops.
I suggest dividing the profile into several pages, like Flickr, so that each page contains 100 photos max.
This modification is absolutely necessary to improve the user's interaction with the site, thank you
another awsome video. Everything explained so clearly.
I'm glad to hear that Richard. There is so much more to cover and what we can do with ControlNet. It's important to understand the basics before you can get creative with it. More to come soon!
Could you do a similar video with the Control Traits when NOT using people as the main subject?
Enjoying your product and the video was easy to follow
Glad to hear that, we appreciate the support!
Fantastic program, fantastic tutorials and tutoring. Thank you for sharing your knowledge, cheers!
Glad you are enjoying PG and the tutorials, we appreciate the support!
Great tutorial, ControlNet seems a little less intimidating now. I like how you break it down, step by step.
Keep 'em coming ❤👌
So glad to hear that. With this video as a basis, the other ones we cover will make more sense. Appreciate the continued support! 👍
Excellent breakdown. Thank you for your efforts. Top notch as always.
Always great to see ya here in the comments bud and really appreciate the support. We're just getting started with CNet, more to come! 👍
This video is worth millions! Because I don't have a premium account, I waste a lot of time generating dozens of images that in the end won't do! And on my own, I've tried every way to use this function to generate different poses, maintaining consistency, and I've never been able to do it! I'll try it now using your tips! Thank you
You're welcome! More improvements are on their way!
thanks your great video! always inspire the learner!
Happy to hear that! You're very welcome!
Another excellent video. Really enjoy your presentation style.
I appreciate that kind sir, thank you! More to come soon. 👍🏼
@@playgroundai looking forward
Thanks for making these videos, they are very helpful and the way you deliver them is very engaging. Even the humour isn't that bad :)
😬😊 I appreciate that! 👍🏼
Thank you sir. I appreciate the video and clarification. I look forward to the Dreambooth filters being available with Control Traits.
You're welcome! Yes the team is working hard on getting that going...I'm looking forward to it too! 😁
Thanks so much for these great tutorials! I had not experimented with the ControlNet image controls but using edge control is already generating images that are amazingly on target with my goals. The image pose seems largely ignored in the prompt, but edge controls works beautifully (and I just started experimenting so much more to go but already blown away with results).
You're very welcome!
Thankyou, very clear explanation as always (where's the 100 emoticon when I need it)
You're very welcome and thank you for the support!
Thank you for this. What do you do with Edge's threshold? From 1-255.
I had a prompt which I wanted to use a Pitbull to replace the cat riding the bike in my image prompt. Can this be done successfully? With what I did, I seemed to get lots of deformities. I used those #'s you used- .6 for Edge & .4 for Depth. Guess practice makes perfect.
Tnx for clear explanation. By my opinion, it's lack of possibility to communication in Playground community. I would be like to have possibility to communicate not only in share on canvas.
The idea to use them together is very useful
For me one reason why not to use MJ
Yes, combining them will usually yield better details and more control, it's not always necessary and dependent on the reference image. So much to show you all still! 🙌
I really like playground AI and the features you guys add. But I wish if you could add more filters/checkpoints like adding some more anime style filters would be great. hope you guys consider this suggestion.
Absolutely! I'm always looking for new ones to suggest to the team, as long as there are no licensing issues. Do you have any favorite's that you like?
@@playgroundaiplease add coloring book filter
@@playgroundai adding some good anime style filters would be really cool.
@@TruthSpeaker-yj6xpyes please! Or at least lineart, then it can be used for a broader range, like tattoos or sketch ❤️🇦🇺
Lovely. Exactly the video I needed to get going with ControlNet. Everything is well explained. TQVM!
Philip T @Playground
You’re welcome and so glad to hear it was helpful! 👍🏼
Very nice!
Appreciate it!
12:43 "Playgraund" yazan bir logo oluşturma
thank you, any idea when playground v2 is out?
You're welcome! Personally I don't have a timeline unfortunately for Playground v2.
I’m deeply pleased with this in-depth guide on depth.
Great to hear bud! Can't wait to see what you do with it!
Is ControlNet feature still working? I only get weird buggy results when i try depth or edge input images.
why I cant generate image in canvas with contol net? its say unknown error and retry
Very good tutorial! As always. ;) Thank you for all of these.
But I have a question regarding the additional control at "Edge". There's a second control slider besides "Control Weight" called "Edge Treshold". Although I have an idea what this is doing, I would like to ask, what is this thing doing exactly? And what are good tresholds to have to get certain results?
Thank you for the great tutorials as always. I'm learning how to use Playgroundai to color my coloring book pages. I've seen Stable Diffusion users do it with 1 page of manga, but I don't know if it's possible to use Playgroundai and Controlnet to do it? I also want to do the reverse: turning pages of colored pictures into coloring book pages. If it's possible, I hope you can guide me.
what does stacking the same trait do?
very cool 👍
Appreciate it!
Can i change poses of an already made image?
Exceptional software. I am really having fun with it! Unbelievable! Even as an artist, I feel like it's cheating big time.
Where are the pose and depth images coming from?
You can use generated images or any stock photo, even your own photos from a phone! 👍
Hi, I don't have "control traits" on board anymore but I've finally found time to experiment with this. Where did the settings for it go? (I have a pro account)
You can only use controlnet with stable diffusion 1.5 at the moment, switch the model on the top right to stable diffusion 1.5. ControlNet for SDXL will be coming soon.
@playgroundai thanks so much. What happened to the playground ai model in the top right? Has it been discontinued? Is there any comms about changes like this I can sign up to? Cheers
Yes discontinued along with SD2.1. You announce things on our Discord and sent out an email to to those that subscribe to our email list.
This explanation/ tutorial is not meant for control traits beginners. i feel the basics is rushed.
what is not clear to you so I could explain better?
Edge threshold was not mentioned@@playgroundai
Thank you for your videos 🙂,
If you please is there a way to generate a story like images something like a small cartoon story or a story board idea ?
Yes there are methods but I will say, generally speaking the nature of stable diffusion is to create random images so achieving 100% consistency if almost impossible. With that being said there are work arounds. Vids to come soon on this! Thank you for the support!
Is there a way in playground to have consistent facial features on various images, for instance as a masked image?
not without training which we don't have yet however if you use a fictitious name or even 2 together in your prompt, you will get a consistent face. I'll be doing a video on this soon.
@@playgroundai Thank you for the reply, much appreciated.
control trait is for paid use only right ? I cannot get the upload image tool for free version
Not at all, it's available for free members as well.
@@playgroundai I see it only work on canvas thank you very much for reply , you already told in video but in the very end that I didn't see it first.
@@playgroundai playground v1 or standard stable diffusion1.5 but for free version has only playground v2 and standard stable diffusion XL , so the free version cannot use because the model isn't avaliable for free user , am I understand it correctly?
@@playgroundai I also cannot adjust the control as pose 0.6 edge 0.4 dept 0.4 , it’s all run together , please let the team check it thank you very much for the team support all issue
12:41 as a logo designer I would love to see more attributes towards this
oooohhhh that would make a great video wouldn't it? Great suggestion! 🙌
first
😆🙌
Dude, how about showing us HOW to bring up the POSE SCREEN??! You just magically had it on your screen already, that's NOT helpful
The graphic was for demo purposes only, if you watched this part 2:19 I explain how that same reference "pose" shows up in the thumbnail of controlnet.
👋
👋🙌
im so paying for a sub once i get some extra bucks.
we appreciate the support!