HOME | DD

Description

The Holi Mandala ContestAs the bleak mantle of winter is lifted, the world is painted in the colours of Spring. Holi, or the Festival of Colours, is a vibrant and joyous celebration of this transformation. ,Holi might be a primarily Indian celebration, but it's observed in many parts of the world in different forms. It celebrates the season of renewal and rebirth, part of the eternal cycle of the universe and life itself. It's a reminder of the unity and continuity in change. Join us in celebrating the spirit of Holi by participating with all the colours in your palette in...The Holi Mandala Contest Create a colourful mandala artwork relating to the theme of Springtime.Showcase your interpretation of the theme through your art.You will be judged on: Creativity, Harmony, Use of Colour, and Interpretation of themeRulesYour mandala must have radial symmetry or some semblance of it.Write a brief description (1-2 sentences) in the Artist's Description of your submission to explain the link between your art and the theme.You can use any medium you like as long as all stock and resources are carefully credited in the artist's comments section of your entry. Please tag your entries with the medium you've used, for e.g. #markers #watercolour #ai etc.If you are using traditional media, make sure you scan or photograph your work clearly so that it is true to the colours you've used.You can submit as many entries as you like, but only one of them may win a prize. Submissions will be classified into two broad categories for this contest: Non-AI and AI.You must mention your deviation is for this contest with a link back to this journal in the artist's comments section of your submission. Remember to include the #holi2024 hashtag.Your work should be created specifically for this contest, which means no work submitted before March 18, 2024.Submit your entry to the contest folder and leave a link in the comments below. Membership to the @Community group is required to submit to the folder. All membership requests are automatically accepted.Dates to RememberSTARTS: March 18, 2024 10:30AM PSTENDS: March 25, 2024 11:59PM PSTMeet your Judges PrizesThere will be 6 winners in total - 3 for each of the two broad categories, Non-AI and AI:1st Place x2 1500 from @team 2000 Fragments from @team Hype badge from @BeckyKidus, @phoenixleo Diamond badge from @AlexanderPaupoff Journal Feature from @Community, @AlexanderPaupoff Llama from @phoenixleo, @Malintra-Shadowmoon, @AlexanderPaupoff 2nd Place x21000 from @team 1200 Fragments from @team Hype badge from @BeckyKidus, @phoenixleo Journal Feature from @Community, @AlexanderPaupoff Llama from @phoenixleo, @Malintra-Shadowmoon, @AlexanderPaupoff 3rd Place x2500 from @team 900 Fragments from @team Hype badge from @BeckyKidus, @phoenixleo Journal Feature from @Community, @AlexanderPaupoff Llama from @phoenixleo, @Malintra-Shadowmoon, @AlexanderPaupoff If you have any queries/doubts, please reach out in the comments below ,#community #journal #news #contest #challenge #colours #holi #dol #mandala #holi2024FAQ What is a mandala? The Sanskrit word 'mandala' means 'disc/ circle', and has various layers of symbolic significance in Eastern belief systems. Simply put, mandalas a symbol for the unity of the cosmos and of the self. It represents the harmony and balance that exists within and without, and governs our existence and the universe itself. Much like seasons, then - each distinct, yet part of an eternal cycle.Cover image credit: Canva
The Holi Mandala Contest
Now, for my AI submission, I chose something *easy and fast*, yet (in my opinion) creative/philosophically valuable - even though it's esthetically nowhere near all the other AI submissions I have seen in the contest
Also you would expect many more segments and variety in a holi mandala - again due to my lack of artistic talent, the source didn't bring any more resemblence of such.
Still, my goal was to use my own abomination uhhh I mean creation as input for the AI and not prompting something completely new.

For this I used a local Stable Diffusion installation, where I used my manually drawn mandala submission for the contest as the base image:

Based on some search, the word mandala can be translated from sanskrit as to "vessel (manda) which contains the essence (la)".
So my prompt for the image-to-image transformation will be exactly this, while I put emphasis on the words "vessel" and "essence (of life)".
Because for me, as I wrote in my hand drawn submission, spring is the revival of what has died in winter.

So to bring life, a spring themed mandala should represent the vessel for the essence of life - at least that was my thought process lol.

I usually never let the AI models decide on their own and just let them generate something by luck.
For this idea I found it interesting to see, what an AI thinks of my image in combination of what it has been taught to be the essence of life.

If you have ever used dreamup, midjourney or any other generator for fun like me, you probably sometimes have written simple prompts like "meaning of life" or whatever.

I remember when I first tested out midjourney when it started up, I prompted "fear" and couldn't sleep for a week.

I still added the most common trigger words most of the AI models out there understand to steer the image into higher quality:
masterpiece, best quality and whatnot.

Not being contempt with the first 5 attempts, I played around with the denoising strenght and parameters like CFG scale.

After that and a bit more of playing around with the prompt, I tried to also add "(radially symmetrical mandala):2" to the prompt, trying to tell the AI to keep it's symmetry and mandala shape.

There are AI models out there that were specifically trained to generate exactly this, however I felt like that would be cheating in my creative experiment, so I stuck to the native stable diffusion model which should at least have been trained on basic concepts.
Meaning - data sets of SD's model have been prepared with tags and intents, while the goal of it's creator was to teach it to generate humans, it *must* have learned essentials of concepts like life, death, emotions.
Give it a million pictures of a sad woman, it will learn how a woman looks - and also how sadness looks - more as a side effect instead of intentionally.
If anything I say makes sense... anyways..

I have put the denoising strenght to 0,55.
This value determines how much the AI respects the original image in the image-to-image generation, while 0 will change nothing and 1 will be fully unrelated to the original.
I did want the AI to decide on its own, but still have it oriented around my original - so I felt like 45% of the original image could stay while the AI has 55% free imaginary play room.

The first 10 or so generations have added a watery look, some resemblences of clouds and kind of a flat depth to the image (which I actually liked very much):

Rinse and repeat and adjust the prompt into the direction you want - as most of my stuff is in anime style, I added "anime style" or "most current anime style" to the prompt, adjusted further, etc.
And the learning in the end is, I have something in common with my stable diffusion installation:
We both don't know for shit what the essence of life is.

Here are some more variations that were generated:

A little important note at this point:Please, do not come to the conclusion that this is a full description of AI image generation processes.
I felt like bringing just a little transparency into the whole topic, while I'm at it.
Please note, that my described process here is just *one* of uncountable other possibilities of using AI tools for image generation!
It usually takes much much more time (at least for me lol), and often manual editting to produce high quality images, while processing images through image-to-image generations is just one of the working steps.

Yes, there are people who simply input a prompt, let it mass generate and then mass submit everything.
Yes, my images might also be very flawed - of course I still have much to learn and it is much fun learning more every day.
And yes, there are accounts on this site who invest more than just some keaboard types into their images - some have more flawed generations while some are just out of my league.
In the end, the topic and possibilites of AI is much vaster than I could describe in a post - while I'm not a "pro" at this, I did earn myself a pretty good knowledge base about the technology behind it all.

If you found this small insight interesting, consider watching me and staying in touch with me.
I love fruitful discussions and I also love contributing to deviantart's community - with and without AI content.

End note:
This all was SO MUCH fun and I took great inspiration from it.

Description

The Holi Mandala ContestAs the bleak mantle of winter is lifted, the world is painted in the colours of Spring. Holi, or the Festival of Colours, is a vibrant and joyous celebration of this transformation. ,Holi might be a primarily Indian celebration, but it's observed in many parts of the world in different forms. It celebrates the season of renewal and rebirth, part of the eternal cycle of the universe and life itself. It's a reminder of the unity and continuity in change. Join us in celebrating the spirit of Holi by participating with all the colours in your palette in...The Holi Mandala Contest Create a colourful mandala artwork relating to the theme of Springtime.Showcase your interpretation of the theme through your art.You will be judged on: Creativity, Harmony, Use of Colour, and Interpretation of themeRulesYour mandala must have radial symmetry or some semblance of it.Write a brief description (1-2 sentences) in the Artist's Description of your submission to explain the link between your art and the theme.You can use any medium you like as long as all stock and resources are carefully credited in the artist's comments section of your entry. Please tag your entries with the medium you've used, for e.g. #markers #watercolour #ai etc.If you are using traditional media, make sure you scan or photograph your work clearly so that it is true to the colours you've used.You can submit as many entries as you like, but only one of them may win a prize. Submissions will be classified into two broad categories for this contest: Non-AI and AI.You must mention your deviation is for this contest with a link back to this journal in the artist's comments section of your submission. Remember to include the #holi2024 hashtag.Your work should be created specifically for this contest, which means no work submitted before March 18, 2024.Submit your entry to the contest folder and leave a link in the comments below. Membership to the @Community group is required to submit to the folder. All membership requests are automatically accepted.Dates to RememberSTARTS: March 18, 2024 10:30AM PSTENDS: March 25, 2024 11:59PM PSTMeet your Judges PrizesThere will be 6 winners in total - 3 for each of the two broad categories, Non-AI and AI:1st Place x2 1500 from @team 2000 Fragments from @team Hype badge from @BeckyKidus, @phoenixleo Diamond badge from @AlexanderPaupoff Journal Feature from @Community, @AlexanderPaupoff Llama from @phoenixleo, @Malintra-Shadowmoon, @AlexanderPaupoff 2nd Place x21000 from @team 1200 Fragments from @team Hype badge from @BeckyKidus, @phoenixleo Journal Feature from @Community, @AlexanderPaupoff Llama from @phoenixleo, @Malintra-Shadowmoon, @AlexanderPaupoff 3rd Place x2500 from @team 900 Fragments from @team Hype badge from @BeckyKidus, @phoenixleo Journal Feature from @Community, @AlexanderPaupoff Llama from @phoenixleo, @Malintra-Shadowmoon, @AlexanderPaupoff If you have any queries/doubts, please reach out in the comments below ,#community #journal #news #contest #challenge #colours #holi #dol #mandala #holi2024FAQ What is a mandala? The Sanskrit word 'mandala' means 'disc/ circle', and has various layers of symbolic significance in Eastern belief systems. Simply put, mandalas a symbol for the unity of the cosmos and of the self. It represents the harmony and balance that exists within and without, and governs our existence and the universe itself. Much like seasons, then - each distinct, yet part of an eternal cycle.Cover image credit: Canva
The Holi Mandala Contest
Now, for my AI submission, I chose something *easy and fast*, yet (in my opinion) creative/philosophically valuable - even though it's esthetically nowhere near all the other AI submissions I have seen in the contest
Also you would expect many more segments and variety in a holi mandala - again due to my lack of artistic talent, the source didn't bring any more resemblence of such.
Still, my goal was to use my own abomination uhhh I mean creation as input for the AI and not prompting something completely new.

For this I used a local Stable Diffusion installation, where I used my manually drawn mandala submission for the contest as the base image:

Based on some search, the word mandala can be translated from sanskrit as to "vessel (manda) which contains the essence (la)".
So my prompt for the image-to-image transformation will be exactly this, while I put emphasis on the words "vessel" and "essence (of life)".
Because for me, as I wrote in my hand drawn submission, spring is the revival of what has died in winter.

So to bring life, a spring themed mandala should represent the vessel for the essence of life - at least that was my thought process lol.

I usually never let the AI models decide on their own and just let them generate something by luck.
For this idea I found it interesting to see, what an AI thinks of my image in combination of what it has been taught to be the essence of life.

If you have ever used dreamup, midjourney or any other generator for fun like me, you probably sometimes have written simple prompts like "meaning of life" or whatever.

I remember when I first tested out midjourney when it started up, I prompted "fear" and couldn't sleep for a week.

I still added the most common trigger words most of the AI models out there understand to steer the image into higher quality:
masterpiece, best quality and whatnot.

Not being contempt with the first 5 attempts, I played around with the denoising strenght and parameters like CFG scale.

After that and a bit more of playing around with the prompt, I tried to also add "(radially symmetrical mandala):2" to the prompt, trying to tell the AI to keep it's symmetry and mandala shape.

There are AI models out there that were specifically trained to generate exactly this, however I felt like that would be cheating in my creative experiment, so I stuck to the native stable diffusion model which should at least have been trained on basic concepts.
Meaning - data sets of SD's model have been prepared with tags and intents, while the goal of it's creator was to teach it to generate humans, it *must* have learned essentials of concepts like life, death, emotions.
Give it a million pictures of a sad woman, it will learn how a woman looks - and also how sadness looks - more as a side effect instead of intentionally.
If anything I say makes sense... anyways..

I have put the denoising strenght to 0,55.
This value determines how much the AI respects the original image in the image-to-image generation, while 0 will change nothing and 1 will be fully unrelated to the original.
I did want the AI to decide on its own, but still have it oriented around my original - so I felt like 45% of the original image could stay while the AI has 55% free imaginary play room.

The first 10 or so generations have added a watery look, some resemblences of clouds and kind of a flat depth to the image (which I actually liked very much):

Rinse and repeat and adjust the prompt into the direction you want - as most of my stuff is in anime style, I added "anime style" or "most current anime style" to the prompt, adjusted further, etc.
And the learning in the end is, I have something in common with my stable diffusion installation:
We both don't know for shit what the essence of life is.

Here are some more variations that were generated:

A little important note at this point:Please, do not come to the conclusion that this is a full description of AI image generation processes.
I felt like bringing just a little transparency into the whole topic, while I'm at it.
Please note, that my described process here is just *one* of uncountable other possibilities of using AI tools for image generation!
It usually takes much much more time (at least for me lol), and often manual editting to produce high quality images, while processing images through image-to-image generations is just one of the working steps.

Yes, there are people who simply input a prompt, let it mass generate and then mass submit everything.
Yes, my images might also be very flawed - of course I still have much to learn and it is much fun learning more every day.
And yes, there are accounts on this site who invest more than just some keaboard types into their images - some have more flawed generations while some are just out of my league.
In the end, the topic and possibilites of AI is much vaster than I could describe in a post - while I'm not a "pro" at this, I did earn myself a pretty good knowledge base about the technology behind it all.

If you found this small insight interesting, consider watching me and staying in touch with me.
I love fruitful discussions and I also love contributing to deviantart's community - with and without AI content.

End note:
This all was SO MUCH fun and I took great inspiration from it.