This fixes all of DALL·E 3's problems...
HTML-код
- Опубликовано: 14 июн 2024
- This is the link to access "The Glibatree Art Designer":
chat.openai.com/g/g-7CKojumSX...
If you also want to Try DodecaLearn, here is the link:
dodecalearn.com/
Keep in mind DodecaLearn is in Early Access, and not really ready for Mobile traffic yet.
This is the link to my Mom's blog post:
glibatree.com/2023/11/09/pain...
Glibatree Discord (Glibacord) - Join the Community!
/ discord
Twitter:
/ glibatree
Instagram:
/ glibatree
TikTok:
/ glibatree
Here is the link to the Midjourney server:
/ discord 
Развлечения
![[The NO Prompt Method] MULTIPLE Consistent Characters with Custom GPT & DALL-E](http://i.ytimg.com/vi/C2DPGWVii_Y/mqdefault.jpg)
Also, Glibatree is FIRE!!!!! I created some really cool art with a prompt guide I purchased. The 4 additional suggestions enabled me to create other variations and even go "off the beaten path" of my original idea - with EPIC results!!!! Thank you so much for this!
Great Work! Excellent video and explanation. Thank you! All the best to you and your mom
You sir are a BEAST for this! Love it!
I’m definitely going to add this GPT to my sidebar. I hate how Dall-E 3 took away 4 images. Thanks for sharing.
Awesome, I hope you enjoy it and make some cool stuff!
Thank you very much. Your GPT is brilliant!!! And it is a terrific tool to learn prompting. Hats off.
Solid prompt work, kudos
Very nice work! Top shelf prompt engineering for sure!
Thanks, Glib! Nice work!
Appreciate it! Thanks for watching!
This is absolutely awesome sir. Really appreciate your help and work.
Totally agree with Bullrun below
Thank you for this! You are a legend.
Thank you for this! Awesome!
Great Work!
Can confirm. This plug-in ROCKS! Thanks so much for sharing.
Amazing, Love your GPT it was what I was looking for
you are amazing. goodness what a legend. Thank you for sharing your high level skills
Stunning
Used your link. The prompt was for a vintage coffee sign with a Corgi and the words "Corgi Coffee" to be spelt exactly. Although the images was amazing, and I would have loved to use them, over and over "Corgi" or "Coffee" were misspelt. I tried 4 prompt attempts, rewording each with more and more specifics. Hopefully the spelling errors can be worked out over time.
You are a G!
Nice! Thanks for sharing!
You bet! I hope you enjoy it.
Really cool art generator. Love this!
Amazing video, thanks a lot, can you share how to enable dall-e to create 4 images per response?
I tried it with a prompt and it did great! But, I'd like it if the user specifies a style, that it uses that style for all 4 images. It might also be helpful if you have the GPT report the GEN_ID for each image generated so the user can easily modify any images made. Thank you!
Hmmm, it should be following the style you ask for it you specify.
You'll definitely get it to do that if you say "Please generate all four in a Macro Photography style" (for example)
Gen_ID is an interesting idea. I'll experiment with how useful I find it
Great, video and great GPT, thank you.
I LOVE YOU, i found yersterday your Custom GPT since it's the number one in the sotre and it's fukken amazing. I will use it as a first option. Thanks so much!
Dodecalearn is awesome, thank you (and your Mom) for that! I have a suggestion: ideally, the number of characters in a row should be between 50 and 60. Lines that are too long make it difficult to move the eyes and tire them, while lines that are too short interrupt the flow of reading. Right now the main text is too wide. Make the content of the text area narrower and center the whole page). Anyway, keep up the good work!
You man rock! I’ve used your GPT recently to recreate some images before using your GPT. Amazing difference. Question: can you add to show the seed number of the images.
So... that tree behind you in the shot is EPIC!!! Where do you get your art printed?
It's a company called canvas champs, they were having a sale and I got all of the ai art behind me for less than $100...
There's a ton of places that do it, I would search for Photo Canvas and try to find the best deal
@@glibatree Thank you!
My biggest problem using Dall E 3 is that if I request a vertical/9:16 image, it's more often than not just a 16:9 image, sideways and mirrored (The text gives the 'mirrored' away).
When you want to change the format of a prompt or request in ChatGPT, starting a new chat session can be helpful. This way, you can clearly specify your new requirements without any confusion from previous interactions. Just copy the prompt you want to use, start a new chat, and paste it there with your new format specifications. If you need any assistance with this process or have specific requests, feel free to ask!
@@Bogi380 hey that's exactly what I need, is it possible to generate stuff using specific aspect ratios ? Like for instance I want to sketch out card back at 0.53:1 aspect ratio, how would one do that ?
Deadly video, awesome.
what interesting is that the image description that Dall-e creates when you click on the image and click the i button, is DIFFERENT then what is out put in the text box. I ran both in Leonardo AI and came back with distinctly different styles. WOW this outputs at 9:16 and if you say for a Reel is totally makes the images awesome.. thank you so much.
A BIG THANK YOU
you are a very very good creator.
I've searched around trying to get my head into this. I've wanted to do this for ages but need a leg up on how to generate the images in my head. You're style is entertaining and really straight forward and you'd be surprised at how many videos bore you to death. Believe me I've watched them ! I'm writing this comment because I think you could grow this channel massively if you had more content and more quick hacks to get to your goal. I'm going to run through your bin and start to learn how to use this. With the very few videos you've created you generated over 39k subscribers. You've said that you want users to understand the deeper parts of Ai - with your understanding you could massively own this subject, just with your style. I do digital marketing and content every day for clients and to get 39k subscribers with 38 videos means you're doing something right. Please be inspired to make more content, as I am inspired to watch your videos.
I clicked on the link to access your Art Designer link but while it asks me to 'join' to access..nothing happens when the ChatGPT 4 opens? could re enphasise "how" to add it to my ChatGPT 4 plus page??
10 outa 10. Love the setup in your gpt
You GPT is the bomb... Do you have one for consistent characters????? On a side note, is it possible to get a new image using a known character or person? Artistic representation, but not as a painting?
You're out here delivering the real and original content. Great work!
Thanks man so kind
wow! i gave your GPT a try and its awesome! very cool.
I was just thinking I wish GPT would make 4 images like MJ. Glad to have found this.
I can't wait till the release the newer version of dall-e
Love the tool! Now we just have to figure out how to get past the 40 requests per hour (limit cap).
So it looks like custom GPTs are no longer offered in Chat GPT 3.5, where in the last video they were.
Niceeeee!! 🔥🔥
I love your course and I have mentioned to my friends that AI will be a boon for art teachers because Art text books are SOOOOOOO expensive. May I suggest that in your art class you can create a section for students to interact more and have AI prompts made as "Funwork" assignments , so students can generate their own images of the art spoken about in the lesson. I don't think it is intuitive enough that a student would automatically think about generating their own art, in the way you mention in your video at 8:00 ish.
I love your website and embrace your moms vision. I ran a private school for a few years, using out side the box teaching techniques, many based upon Marva Collins style of allowing each child to work at their own pace.. it worked like a charm.. and most of the children from kindergarten were reading far above their age and grade even by the end of their 1st year and they excelled at standardized tests and scored far above their grade levels if they had been straddled with being in Public schools. Public schools were created to pump out compliant factory workers.. not help people excel .. it was made to keep people in LINE.. hence in your Grade.. if we don't have grades then kids don't have labels, and can excel beyond expectations. More power to you both.. I am rooting for your success.
Thank you. I’ve been using your custom instructions. Now this is even better.
Wonderful to hear!
wow this seems like a really cool tool
Yay! I hope you enjoy it and make some cool stuff 🤩
Please consider adding 'show seed' to the generation. Then we can 'use seed' to iterate on already created images. Thanks!
Is there a video showing how you configured your GPT?
Which planet are you from? Your mom has to be so proud! Amazing work. Awesome.
Awesome.. I know you love Midjourney but can you have it set up so that the user can ask for a Leonardo AI output instead? Not a big deal.. but would be cool if you have a list of things you want to add. Thanks for this. I am excited for all the helpful GPTs coming out.. my side bar is jammed.. lol
The reason I mention Leonardo AI is because it is super duper easy to take the square images made with Dall-e and put them into the Leonardo AI outpainter and create larger either landscape or video portrait sized images.. and Leonardo AI will also add image to video soon .. and it is cheaper for people to use.. so just a thought.
When you generate 4 images does that use up more of your allowance or now
Bro bro... I just wanna discuss this stuff with somebody face to face. I've developed a full graphic novel layout and character generation with consistency!
I never thought it would be possible
That’s awesome! I wish I could pick your brain lol
Holy cow! You gotta share that with the world.
This time next year we'll be millionaires Rodney! Awesome work 🙏👍
Thanks for sharing man. Results are great but when I use the numeric value to kick start the next round of 4 images (from the 4 suggestions), I only get one image. Are you consistently getting groups of 4?
I have the same issue
It consistently works for me at the beginning of a conversation, but as the conversation gets longer it sometimes stops following all my instructions.
Try adding this to one of your prompts, and it should get back on track:
"Also (going forward) please remember, whenever I ask you to generate something - am looking for you to create for interesting variations. I always want to see as many version of my ideas as possible."
Dunno why but it always gives an error when trying to generate the Dall-E images.. The Midjourney prompts do come through do..
Wow! a mind blowing tutorial. But why all videos that demonstrate Dalle-3 the generate only Square dimension images? why you don't demonstrate to us of changing the aspect ratios?
Will one prompt (resulting in 4 images) equal 1/50 GPT4 prompts/ 3 hour or does it count for each image?
You might hit a secondary image generation rate limit behind the scene but one request will only ever count as 1 of your 50 images per three hours
@@glibatree thanks. I will likely not hit it anymore as I’m using leonardo ai for the most part because of their API.
OMG where is the future going with this?
does this work well with consistent characters ?
hey I just tried you gpt it's good 😊. I just wish that dall-e 3 would have a better time with making eyes look better and make small details look better
It keeps resetting the resolution to 1:1 no matter what I try and tell it to do. At most I can get 1 image in a different size and unless I specify every time, it resets to 1:1
If your trying to get widescreen or portrait, that is probably a problem with my instructions - I'd be curious to see what you are asking exactly - cause I bet I can improve how I control this.
For now, it works for me if I add "I would like all 4 images in a 16:9 aspect ratio." to the end of my request. So maybe try that in the meantime.
@@glibatree we added to every prompt to work around.
In you original custom instructions you had a "use 1:1" rule that we ended up changing.
@@glibatree normally we ask gpt to always create images in 16:9 and have to 'reset' the rule after 10 images or so.
"Please create all images in 16:9 for this conversation" usually works. Does not with your rules overwriting
@@AbsoluteBestSlotJackpots I see what you mean. I can work on making the defaults not quite so rigid. I does seem like it would lose track during long conversations.
@@glibatree i like how you had some unique blend ideas via your mom. I am trying to get more exotic blends of styles. Thank You
How do I get it to do these CGI quality images that you are creating?
All the images in this video I generated using the GPT either directly or by pasting the Midjourney command it generated into Midjourney
How can I find the gen_id for my images? Can you add that your amazing plugin.
Noob Dalle-E user here… is it possible to tweak this so that it generates the four images in one style that I request in my initial prompt? If so, how? Thanks, nice work 🙏👍
I hope you see this message soon but glibatree and midjournay is broken. It was working last night but now when I copy a prompt I am getting this message: Invalid parameter
Unrecognized parameter(s): glibatree, prompt,, hd, quality,, natural, style
/imagine A charcoal sketch of an African American woman with stylish locs, dressed in a teal business suit, deeply engrossed in her work at a minimalist black desk. She is scrutinizing a document with a yellow highlighter in hand, the room illuminated by a soft, diffused sunlight from behind. Drawn with:charcoal shading techniques, high contrast between light and shadow, emphasis on facial expressions --ar 1:1, glibatree prompt, hd quality, natural style
Yeah, it looks like the GPT got confused by the Midjourney format. This should never happen at the beginning of a conversation, but as your context gets longer and longer it sometimes gets confused.
You of course can start a new one, but sometimes you like the results after a lot of tweaking. So to fix it and get the bot back on track, try sending it this message:
"Remember that the --ar 1:1 needs to go at the very end of the prompt. Please regenerate all four Midjourney commands with this in mind"
if only i could find a character sheet GPT
I don't know but it is not working since few days.. not able to generate images in chatgpt. It is not able to generate images
I'd love to try it... If I had access to GPTs 😢
Any day now, I'm sure, haha
```Midjourney
/imagine prompt:A digital comic-style illustration of a professor giving a lecture in a crowded university classroom, the professor animatedly explaining a complex topic, surrounded by eager, diverse students of various descents and genders. The classroom is bustling, filled with desks and educational posters. The scene is vibrant, capturing the energy of learning. Drawn with: Quentin Blake-inspired loose, expressive line work, exaggerated character expressions, lively and bright color palette, glibatree prompt, hd quality, natural look --ar 1:1
```
I'm unable to generate images based on the description you provided because it doesn't align with our content policy. Let's try a different approach or concept for your image request. Please provide a new idea or adjust the current one, and I'll be happy to help create a suitable image prompt.
```Midjourney
/imagine prompt:A digital comic-style illustration of a professor giving a lecture in a university setting, depicted in a style inspired by Quentin Blake. The professor, engaged and expressive, stands in front of a blackboard filled with diagrams and notes. The university ambiance is captured with bookshelves and academic decor in the background. The scene is lively, reflecting the dynamic nature of education. Drawn with: loose, expressive line work, vibrant color palette, playful character design, glibatree prompt, hd quality, natural look --ar 1:1
```
I'm unable to generate images based on your request because it doesn't align with our content policy. To create an image that fits within our guidelines, we need to adjust the style reference to Quentin Blake, as we cannot replicate the style of artists whose latest work was created after 1912. Let's try a different approach or concept that adheres to these guidelines. Please provide a new idea or adjust the current one, and I'll be happy to help create a suitable image prompt.
That's great but can we have a transcript of the GPT instructions so that we can create and modify our own GPT. E.g. I might want a GPT that will create images to put on t-shirts so I need to add specific instructions about having an easily removable background, soft irregular margins so the image blend with the fabric without square edges etc.
This GPT should be able to support custom instructions - but I didn't write this in a very customizable way (I used every character I had haha).
If you want to use my GPT to get that sort of image - I would try adding exactly what you said to your own custom instructions and I bet The Glibatree Art Designer will give you what you're looking for without worrying about rewriting it.
@@glibatree I've got my own GPT that I'm working on, is there any chance you could let me know what text you put in to get it to do 4 images, so that I can modify my own to do that?
@@vodkaman1970just tell it to repeat n times !
Hi I love Your videos
I been using Chat GPT Dall - E 3 to create images
I have a question about different ways of creating content
I am looking to create the showing process of my AI ART being created over time in real-time so I can record the process and post a 10-15 sec videos of the process on social media but I do not know how? Thanks God Bless
Link please for your custom gpt
how about a walkthru of how you built the GPTs
Good idea for another video. GPTs seem to have a lot of intricacies if you want to make them cool.
How do you make the prompt default to --ar 16:9?
Just mention you want the images to be widescreen or 16:9 in your message and the GPT will make the commands for you
♡♡ graet
👋
👋👋
I keep getting this error generating images: "It seems there was an error in generating the image. Let me try that again": it also states using unknown plugin. I guess I will try again later. Even when using Dall-E 3, it is saying it can't generate images.
I get the same everywhere; seems like image generation through all GPTs is down at the moment (including OpenAI's default GPTs)...
Hopefully this video didn't cause too much traffic for them to handel. The tool is so good I would hate if they nerfed it's capability.
@pjohn73 Looks like it is back working!
Is the content created with your chat GPT private?
I have no access whatsoever to see it; not sure if OpenAI has a backdoor or whatnot - but it is as private as anything else from ChatGPT
Hmm, we glossed over the build instructions for your custom GPT no? Maybe I have to watch it again. :) Maybe your large copy/paste from the art styles would be something you put in the configure section of your custom build GPT? I would have liked to see how your actual GPT was built. Cheers! Good luck to you and Mom!
Ahhh, so it's a secret sauce. I notice someone else mentions to add instructions to not show your custom instructions...I get it now.
Yeah, I think part of it is I want to see how the GPT Store/community revenue sharing ends up working before I become the catalyst for thousands of clones.
I also think there's value in making this one as good as I can, instead of sharing the prompt only to have ChatGPT update and break the prompt again
I would definitely make a video about how I make GPTs in the future though, cause it's an interesting process that I think lots of people would love to apply
Cool. Ya, it makes sense. Thanks!@@glibatree
@@glibatree The GPT Store thing though confuses me, as currently only other GPT Plus users can use Custom GPTs. Would be of much more value, if the Custom GPT links created could then be accessed by anyone (even non-OpenAI subscribers).
By far the best GPT i've used!
Thanks man so glad you enjoy it!
I think its crashing. I can generate in normal gpt but this does not want to work.
Given others have it working i wonder if its the generation of 4 in one hit causing it
I get the same everywhere but ChatGPT proper; seems like image generation through all GPTs is down at the moment (including OpenAI's default GPTs like "DALL E" and "GPT
Cosmic Dream")...
Hopefully this video didn't cause too much traffic for them to handle. The tool is so good I would hate if they nerfed it's capability.
We may just have to give it some time for it to come back.
@flightsimtoday Looks like it is back working!
you look and also sound like Ray Romano :)
Why would I want the prompt to use on a inferior AI?
You might want to add refences to your Learning app because a lot of those pictures generated in your post look nothing like the art styles depicted in the Vatican museum that has a great deal of Egyptian, Greek. Byzantine are etc. Other than that nice GPT
Haha thanks. Yeah that post was my mom's first ever time generating AI Art at all. it was more that she had a good time fulfilling her vision - not necessarily that they are perfect or Historical. If we add images to the learning app would likely use real references from primary sources - and have GPT-4 V analyze, describe, and teach about them.
@@glibatree I am working on few myself in the education space so I have been formatting my text files for the GPT so it works the better.
@@southcoastinventors6583 neat that sounds very cool
Hey nice work! Just want to ask that are you interested in our new ai generation product? Maybe I can send you beta version to you~
Crashes often when it gets to the 3rd image
I find that has more to do with ChatGPT's bandwidth than my prompts, maybe try again now or after a few hours.
Dude, big improvement in image quality. Hope you make lots of money off this.
Love the effort. Getting amazing results. But the numeric thing is kind of random, sometimes it generates one image for me and some times 4.
Edit: Now the regular prompt give me just 2 images instead of four if I use the same chat. If I use a new chat then I still get 4 images.
Seems like the conversation you're in has fallen into a rhythm that and it's not really following my instructions either.
Try adding this to your next prompt:
"Also (going forward) please remember, whenever I ask you to generate something - am looking for you to create for interesting variations. I always want to see as many version of my ideas as possible."
This should probably get you back going in the chat you are in.
broken
Seems to be working on my end. Could be that OpenAI had a hiccup and is back now, or perhaps your running into a specific problem and maybe someone can help.
If you are still having issues, try posting some details in my discord and someone will probably have an idea.
Please give instructions for your gpt to not give your custom instructions. So people can’t reverse engineer your gpt
Haha very wise. I'm also still in a mode where I'd want to test any updates before I roll out this to everyone - because these can be a little finicky and I wouldn't want people getting worse art because I was being defensive over some concept of secret sauce lol
But it is a good idea, I'll try to add something along these lines in the next couple days.
Thank you for some of the sauce lol! I’m working on a children’s book scenes however I want to keep the background the same overall but with added elements and character’s to be different. Any suggestions with prompt images?
•//° I believe we are meant to be like Jesus in our hearts and not in our flesh. But be careful of AI, for it is just our flesh and that is it. It knows only things of the flesh (our fleshly desires) and cannot comprehend things of the spirit such as peace of heart (which comes from obeying God's Word). Whereas we are a spirit and we have a soul but live in the body (in the flesh). When you go to bed it is your flesh that sleeps but your spirit never sleeps (otherwise you have died physically) that is why you have dreams. More so, true love that endures and last is a thing of the heart (when I say 'heart', I mean 'spirit'). But fake love, pretentious love, love with expectations, love for classic reasons, love for material reasons and love for selfish reasons that is a thing of our flesh. In the beginning God said let us make man in our own image, according to our likeness. Take note, God is Spirit and God is Love. As Love He is the source of it. We also know that God is Omnipotent, for He creates out of nothing and He has no beginning and has no end. That means, our love is but a shadow of God's Love. True love looks around to see who is in need of your help, your smile, your possessions, your money, your strength, your quality time. Love forgives and forgets. Love wants for others what it wants for itself. Take note, true love works in conjunction with other spiritual forces such as patience and faith (in the finished work of our Lord and Savior, Jesus Christ, rather than in what man has done such as science, technology and organizations which won't last forever). To avoid sin and error which leads to the death of our body and also our spirit in hell fire, we should let the Word of God be the standard of our lives not AI. If not, God will let us face AI on our own and it will cast the truth down to the ground, it will be the cause of so much destruction like never seen before, it will deceive many and take many captive in order to enslave them into worshipping it and abiding in lawlessness. We can only destroy ourselves but with God all things are possible. God knows us better because He is our Creater and He knows our beginning and our end. Our prove text is taken from the book of John 5:31-44, 2 Thessalonians 2:1-12, Daniel 7-9, Revelation 13-15, Matthew 24-25 and Luke 21. Let us watch and pray... God bless you as you share this message to others.
Very cool
Dall e 3 good only deatils..but realistic and Vision dont...