Personally it is the wildness that has an effect on me and the painted like quality.
Almost like a dream it creates strong eerie feelings but in a good way. Craiyon and it seems DALL-E 2 are more literally making an image without that effect on the subconsciousness, at least in my case.
3 Likes
Yeah, that is one effect everyone considers the most I guess. This DALL-E needs a really good prompt if you want it how you want it to be.
My first try of ‘river of fire’ prompt from previous thread doesn’t look anything like the one created by Midjourney.
But this is if I modify the prompt to ‘a realistic photo of a river that is on fire’
7 Likes
Mrs. Chou, street lady of your nightmares. She comes up a lot in mnemonics (there’s even one where she kills somebody: 弔), so this is actually a really nice mental image to file away for later. So cool.
4 Likes
Do you want to commit and make this script for all the mnemonics or maybe just for fun only for this thread? I would love to help you if you don’t mind as I’m free this summer holiday 
One thing comes to mind is that maybe we create a public repository first contains all the mnemonics sort by levels, then people can submit their image to the repo. The image for the script is chosen based on likes and comments.
I personally chose Midjourney for default model as it is wilder and it has seed feature as explains here. This case we can maintain the look for Jourm, Ken the Samurai, Mrs Chou, and so on.
3 Likes
I don’t see how the explanations of seeds can help to keep images consistent.
2 Likes
As explained in the doc, seed can “sometimes help keep things more steady / reproducible between generations.” But I guess it only works where we want a same prompt to be reproducible, but it won’t work in case we want the same face of Mrs Chou in different prompts.
These are my first and second try for ‘curly haired Asian woman with red eyes and angry’ without seed.
And these are two prompts with an additional seed command, ‘curly haired Asian woman with red eyes and angry --seed 1249’. As you can see, both look similar.
6 Likes
Ok, that’s plausible. I thought there is a function to use a character from another user in a new image.
It would be useful to be able to define elements, save them and reuse them. I found that it is not possible to use a lot of keywords.
2 Likes
Another option is to create a discord server for this thread and invite Midjourney bot to the server. In that case, we can just vote using emoji and make it favourited for the most liked, just like what they do in the official Midjourney discord server.
One question though, is there any already available list for all the mnemonics in WK sorted by levels?
2 Likes
Sorry, I used up all my free trials with a repetitive futile try to visualize 蛇.
I plan to make the cheapest subscription but to be honest I’ve never used discord a lot and don’t like subscriptions in general so it might take some time.
1 Like
No worries! I do have a subscription, so feel free to request 
2 Likes
I thought it looked like Liz from Spiderman Homecoming…
2 Likes
No! I cannot resist! 
You hoist your towel above your table, drop into the water and sail away. You turned the towel into an actual sail and you’re balancing on top of the table, which is now a boat.
Is it difficult to make a subscription from within discord btw?
1 Like
I can see it!
2 Likes
After several trials of prompting lol, it is difficult for the model to get the idea literally but I guess this can resemble the prompt since I guess it is the mnemonic for sail.
But I like this better from DALL-E 
7 Likes
Thank you! That’s cute!
How I want to be on this boat now 
4 Likes
Not at all ^^ They have a website and it takes about 2 min.
1 Like
Sameeee!!!
Let’s all try to go in a boat at least once till the end of the year
3 Likes
I would like to see this one: “There’s a samurai with (just) his forehead and legs. Everything in between is missing. What happened to everything in between? He was low on money, so he had to sell it all, leaving him with only the top and the bottom.”
2 Likes
