I strongly advise on the midjourney V5 and above to put the image link after the description, it seems to have a better understanding, then add always the max image weight (2.0) at the end of your prompts if you want results close to your original image Ex: your description, the link then --iw 2.0
I don't know how you guys feel about this, but my conclusion so far is that in V5, the improvements in realistic rendering of photographs have come at the expense of creativity. Artistic styles are no longer weighted in the images like before, and the specification of an artist in concrete image descriptions is no longer considered as much as it was in V4. Sometimes the artist doesn't get any attention at all and the motives look quite similar, even if you specify a concise artist as the style. For a series of concept art images with a comic-like touch that looked really great in V4, I get quite disappointing results in V5, no matter how I rephrase the prompts. What are your experiences?
I find the same. Have you tried adjusting the -stylize in your prompts? I am resting it out today. I find -stylize 750 brings some of it back. But not as much as v4 (so far)
@@Carskinsmollie Good point! They are tweaking it all the time :). I believe they are now running a faster version (V5) and they put the original alpha to --V5 A :)
The tutorial is great, BUT if we use styles from a website, it doesn't actually apply the style to our photo, but rather mixes it with the 4 photos that represent the artist's style. If you need to stylize a man and you choose a style where women are depicted, then it will mix the woman and man from the photo depending on the weight of the photo and the style. To make the tutorial work, you need to find an artist who has drawn something similar to your photo and preferably of the same gender.
Fantastic tutorial! The Midjourney Reference List you put together is incredible! But is there a way to download that locally? It takes too long for the images to load. Tried downloading directly from Google but it just times out.
I've uploaded one of my own images and hit enter. When I right click on the image I do NOT get the "copy link" option. I have "copy message link". When I try that option I always get "Could not validate this link. Please try again later." What am I missing??
Great video, wombo does this very same thing on a nice mobile/ipad app. I am new to AI but I am wondering why midjourney is still popular despite being on a wonky discord server and need to learn all these complex commands
Only reason I can think of is that the results on mj are better than all it’s competitors. If another app catches up, mj is toast. I’ll try wombo tomorrow and see how it goes :)
hey awesome video, helped me a lot, I just can't force my discord to give me a remix prompt it just start creating variants, is it a mac feature only or I mist something?
Hmm, no remix option? It is not a mac feature as far as I know. Type in /settings and then check if "remix mode" is selected (it should be green if it is on). Let me know if that works :)
Kind of newbie here, how mid journey knows the artist names? And their styles? Is the midjourney trained on the images of these artists? Let's say I start making many images from today in my own style and then you like my style, then you won't be able to write my name in the prompt to get the results like right? Please enlighten me how does this connection work. Thanks
Wow! Awesome for your great helpful sharing . It encourages me to watch and learn more of what you share and do Midjourney practice and creativity. Great work, great sharing. Yes, this time it is wow for me. It is what I am looking for and maybe beyond❤😊.❤🎉😊 thank you so much.
Following the steps but after uploading the reference image into the chat, when I right click on it I only get eh option to copy the "message link" not simply "copy link" which seems to holt the process, any ideas?
Great info and well done. Thank you! I do have a question for you. If I am using a image that has no character in it, lets say its a picture of a vehicle. Instead of typing "character" can I type "vehicle"? Also how do I get it to match my vehicle image, I want the same exact vehicle just a different style, is that possible?
Damn I'm coming up with so crazy/scary/weird and some just ...... not right pictures. The close up portraits work the best. When I start adding full body and background items then it turns weird.
Portraits definitely work best. Especially because they tend not to have hands and fingers in them, and those are still kinda wonky (though not as bad as previous generations). Good point.
@@Martinwhynot Preset your character & location so it knows and you can reuse it, I mean it shows that it knows celebrities but it can't save your own characters, I think until GPT4 rolls out with the vision side where Ai can see what it's doing this won't be as possible
I keep getting error messages . "Banned prompt detected Sorry! Our AI moderator thinks this prompt is probably against our community standards." and Invalid link Could not validate this link. Please try again later.
@@kingy-ai Also I love your explanations, details, and content you've been creating. You've really helped enhance my Midjourney crafting skills. Thank you!!
Are you a subscriber? :). Here is the link to the image: media.discordapp.net/attachments/1004210781786751108/1086181709831221288/Big_C_attractive_man_character_by_Becky_Cloonan_c4b5187a-fdd8-4d7c-8ea5-f2f4c889f224.png?width=1340&height=1340 attractive man, character by Becky Cloonan::1, s.mj.run/DnWOjCF7tmo --iw 0.75 --v 5
Thanks so much! I've been looking all over for cartoon styles for my images, as I want to make comic books and graphic novels of my characters. Do you have a video for varying facial expressions, poses, etc while keeping the same character "look"?
I don't have one yet, but this is something I am looking to do asap. I am going to research and test this tonight, see what I can come up with! Great idea! :) Thank you! :)
@@kingy-ai This would really be an amazing video. For reference, a self-help TH-camr called Hamza has really achieved this with his thumbnails. You can clearly see it's him in all of his thumbnails, so the person doing the prompts is able to use the same character doing different things based on the video's subject. I've been trying to achieve this with my own images but with no success. It would be amazing if you did a tutorial about this.
Still didnt find what i was looking for, you said you will go through till the abstract part, i do not see any abstracts. End up so much talking about the other artists.
And when I copy the message link, and paste it into the bot, I don't get a second image - just the link spelled out. If I click on that, it just jumps up to the first image I dragged in.
I'm annoyed. At myself of course. I have a personal photo of a friend, that I'm trying to turn into a more cartoony art style. No matter what artist I choose the image keeps coming out realistic. Help?
I don't know why but it keeps doing me the illustrator face instead of my media. Anyone has this problem? Thank you for your work i'm a Newbie and this is helping a lot!
Bro is use stable diffusion 2.1 in Leonardo ai in custom model and i am facing a problem that my promt are generating bad eye and other pose with many problems😢😢
You got a sub in me and you are my go to everything midjourney. I keep trying to use a video game screen shot and I want to either make it realistic looking or my comic book. All I get is imagine link is invalid. I upload it just like you say. I tried 4 different pics. Please help me what am I doing wrong? Do you have a video on upload game art? I am going to look. This was informative and I pray it helps
I tried this several times and I can never get it to look like the artist in the original photo. I've tried --iw 0.5 and still it seems to combine the image with the illustrators model.
pls help..I copy the link enter in image prompt and do everything whatever is required ..I copy image link after opening in browser as by right clicking on the image it show copy message link with a link symbol a..I tried this copy message link as well bt after hitting enter the output is nowhere close to the reference image in fact it shows a different image only?? I have tried it multiple times but not working. Is it something to do with the settings?
@alinaminute thanx alot for this tutorial I have questions please Now if i generates ai art and i like it very much So i want to replace the face in that ai art to my face I tried alot but its always doesn't look like me at all🤦
Hey there. Why cant i generate a original prompt today, even as a trial and with the /imagine command on, Midjourney tells constantly due to excess demand they cannot provide a trial version, to try in next day, or subscrive. What the hell?! Its AI! Thanks.
Midjourney is definitely a discord app. That said, I just looked on the Apple App store and there is a free Midjourney "type" app - but it looks suspicious to me. It's called "Midjourney AI · Art Generator" on the app store. I haven't tried it as it it looks sus.
I think I'm doing something wrong because I never get images that look anywhere near my pictures. I'm a man with a beard and I get women images, per example...
Same for me. I'm trying to change the image style or the place but trying to keep my identity, my face and I can´t get it. Let me know if you find the solution please!
@@kingy-ai I set --iw 2.0 but the face was completely different. **sigh** I wanted to show a new cool technology to my kid :) I do enjoy using chat-gpt though.
I strongly advise on the midjourney V5 and above to put the image link after the description, it seems to have a better understanding, then add always the max image weight (2.0) at the end of your prompts if you want results close to your original image
Ex: your description, the link then
--iw 2.0
Great point! Thank you for sharing :).
sample?
Which is your favourite Cartoon / Avatar? Mine is the Messi one!
👋
Hyper Realism photos
That Ronaldo one is SICK!
I love it when things on my screen don't seem to work the way yours does. I'll figure it out. Good video.
I don't know how you guys feel about this, but my conclusion so far is that in V5, the improvements in realistic rendering of photographs have come at the expense of creativity. Artistic styles are no longer weighted in the images like before, and the specification of an artist in concrete image descriptions is no longer considered as much as it was in V4. Sometimes the artist doesn't get any attention at all and the motives look quite similar, even if you specify a concise artist as the style.
For a series of concept art images with a comic-like touch that looked really great in V4, I get quite disappointing results in V5, no matter how I rephrase the prompts.
What are your experiences?
I find the same. Have you tried adjusting the -stylize in your prompts? I am resting it out today. I find -stylize 750 brings some of it back. But not as much as v4 (so far)
The v5 model is still in alpha, the devs have come out and said that when the official v5 comes out it will be more stylistic
@@Carskinsmollie lets hope so!
@@Carskinsmollie Good point! They are tweaking it all the time :). I believe they are now running a faster version (V5) and they put the original alpha to --V5 A :)
....mmm, let see, " I like this guy, mmm...Glen Keane! " hahahahahahaahhaaha , I guess you know who that guy is right? XDDDDDDD. Thank for the video!
The tutorial is great, BUT if we use styles from a website, it doesn't actually apply the style to our photo, but rather mixes it with the 4 photos that represent the artist's style. If you need to stylize a man and you choose a style where women are depicted, then it will mix the woman and man from the photo depending on the weight of the photo and the style.
To make the tutorial work, you need to find an artist who has drawn something similar to your photo and preferably of the same gender.
This is fascinating. I need to watch it a few more times to understand what you did!
Hi
My name is Wael from Egypt, i love your videos, it's amazing, short & clear, thanks for this content
Thank you Wael!!! What part of Egypt are you from? :))
@@kingy-ai I'm from the capital Cairo
Hell yeah, straight to it, no bs. Thank you
I got you! :)
Mines doesn't show a copy link tab it only shows copy message link, can you help with this?
Same. So frustrating. I’ve tried it on mobile and pc.
@@shekinahglory604 I figured it out. You have to do it in the app.
What app ? On computer or phone ? Because I use my phone and go into discord app. Can you help me
Thank you very much for the reference links, those are terrific resources.
It’s my pleasure!! :)
Thank you! I was exactly looking for this.
I got you!
Excellent job, my friend! I'm pretty stingy with "thumbs up" but you earned this one! 😁
Thank you my man!!
Woah love your channel man ❤
Just what I needed 😊
Hell ya! Thank you Yusuf! My man!
What do you think Aaron McGruder, Alex Ross, Becky Cloonan, Neal Adams, Gil Elvgren, etc think of this?
Ask them...lol
@@OneYulaw Oh I will, but that's not the point. The question is for @AIinaminute to put himself in their shoes.
Very helpful. awesome. keep up the good work.
Awesome! Thanks for prompt docs
Very nice Reference list, thank you very much for sharing~~~
AMAZING guide and thanks for the reference links!!!
I got you :)))
Very helpful thank you very much
Fantastic tutorial! The Midjourney Reference List you put together is incredible! But is there a way to download that locally? It takes too long for the images to load. Tried downloading directly from Google but it just times out.
Thanks! So helpful.
I signed the basic plan for 200 images per month. Does each variation and upscale spend my credit?
Hmm. I believe it spends credits. But at half the rate of v4 due to new efficiencies. :)
I've uploaded one of my own images and hit enter. When I right click on the image I do NOT get the "copy link" option. I have "copy message link". When I try that option I always get "Could not validate this link. Please try again later." What am I missing??
this! did you find a solution?
Thank you for sharing!
I got you!
Thats mindblowing :P 😮😮😮 thank you 😊❤
It really is! Glad you liked it! :)
How are you getting the remix prompt dialog? Discord just doesnt popup a remix prompt for me when I try the variations.
hey hey! type in /settings and check if "remix mode" is selected. Make sure it is green :). that should do the trick :)
Great video, wombo does this very same thing on a nice mobile/ipad app. I am new to AI but I am wondering why midjourney is still popular despite being on a wonky discord server and need to learn all these complex commands
Only reason I can think of is that the results on mj are better than all it’s competitors. If another app catches up, mj is toast. I’ll try wombo tomorrow and see how it goes :)
Wombo is/was pretty basic and silly app. I had it on my phone once but recently it says it's not available on the apple app store in my region.
hey awesome video, helped me a lot, I just can't force my discord to give me a remix prompt it just start creating variants, is it a mac feature only or I mist something?
Hmm, no remix option? It is not a mac feature as far as I know. Type in /settings and then check if "remix mode" is selected (it should be green if it is on). Let me know if that works :)
@@kingy-ai oh hehe, of course it's in the settings, thank You and best regards
Amazing Information! Thanks so much for sharing😊
I am glad you like it! Much appreciated :)
Kind of newbie here, how mid journey knows the artist names? And their styles? Is the midjourney trained on the images of these artists? Let's say I start making many images from today in my own style and then you like my style, then you won't be able to write my name in the prompt to get the results like right? Please enlighten me how does this connection work. Thanks
Wow! Awesome for your great helpful sharing . It encourages me to watch and learn more of what you share and do Midjourney practice and creativity. Great work, great sharing. Yes, this time it is wow for me. It is what I am looking for and maybe beyond❤😊.❤🎉😊 thank you so much.
It is 100% my pleasure!!! Glad it helped you! :)
Very helpful video, thank you
Glad it was helpful! My pleasure! :)
For me it loses most of the original person and puts in the charsteristics of lets say Robert Crumbs
Nobody is going to talk about him doing this at 2:00am
dedication 📈📈
Damn rights! This video took me 10 hours to make lol. It has been Totally worth it though!!! :)
You forgot to mention turning on remix mode /prefer remix
Following the steps but after uploading the reference image into the chat, when I right click on it I only get eh option to copy the "message link" not simply "copy link" which seems to holt the process, any ideas?
look down further the right click menu. Unless it is an unsupported format?
@@AdaptiveDevelopementPlatform thanks, I solved it by opening the image in its own browsers from the chat and then saving that address
thank u that was useful
My pleasure :))))
why cant i right click on the image after uploading. hmmm.
Ronaldo SUIIIPERMAN is dope
Hell ya. That’s a seriously good piece of A.I. !!!
But you can't have custom image weights with Version 4. I get "Invalid Parameter" for that reason.
Custom image weights do not work with v4. Very good point.
I am getting this too! I tried putting --v4, but says doesn't recognize that. What are we doing wrong? How come it works for the video? halp,
Thanks!
❤️❤️❤️
Great info and well done. Thank you! I do have a question for you. If I am using a image that has no character in it, lets say its a picture of a vehicle. Instead of typing "character" can I type "vehicle"? Also how do I get it to match my vehicle image, I want the same exact vehicle just a different style, is that possible?
When I right click the image, I see "copy message link", but not "copy link".
Fantastic 😂
Damn I'm coming up with so crazy/scary/weird and some just ...... not right pictures.
The close up portraits work the best. When I start adding full body and background items then it turns weird.
Portraits definitely work best. Especially because they tend not to have hands and fingers in them, and those are still kinda wonky (though not as bad as previous generations). Good point.
how do I use my own images instead of fr the website library?
I'm subbing just joined Mid journey V5 my problem is when will they allow you to preset your own characters, Locations & objects
I may be misunderstanding your point but I think you currently can if you have your own work as a JPEG file and follow these instructions.
@@Martinwhynot Preset your character & location so it knows and you can reuse it, I mean it shows that it knows celebrities but it can't save your own characters, I think until GPT4 rolls out with the vision side where Ai can see what it's doing this won't be as possible
@Art Official Imagery can you save your character & have it remember
Great video! what are the full scales of artist style and original image, 0 - 2.5? confused on the different options for both numbers
I believe for image weight he mentioned that 2.0 is the highest setting. Not sure for the other option.
I keep getting error messages . "Banned prompt detected
Sorry! Our AI moderator thinks this prompt is probably against our community standards." and Invalid link
Could not validate this link. Please try again later.
when I hit V3 it doesn't bring up a remix prompt - is that a setting ?
how do i post in solo server and not with other users, they post so much, Mine gets lost and have to scroll up so much everytime
Hey, have you tried using IMI prompt for selecting genres and looking at a library of styles?
I have not tested that no. It looks interesting so far (kinda like Noonshot, but with more features)? How do you find it?
@@kingy-ai Also I love your explanations, details, and content you've been creating. You've really helped enhance my Midjourney crafting skills. Thank you!!
Please, I need the "Messi in the Real Madrid" prompt hahaha
Are you a subscriber? :). Here is the link to the image: media.discordapp.net/attachments/1004210781786751108/1086181709831221288/Big_C_attractive_man_character_by_Becky_Cloonan_c4b5187a-fdd8-4d7c-8ea5-f2f4c889f224.png?width=1340&height=1340
attractive man, character by Becky Cloonan::1, s.mj.run/DnWOjCF7tmo --iw 0.75 --v 5
Thanks so much! I've been looking all over for cartoon styles for my images, as I want to make comic books and graphic novels of my characters. Do you have a video for varying facial expressions, poses, etc while keeping the same character "look"?
I don't have one yet, but this is something I am looking to do asap. I am going to research and test this tonight, see what I can come up with! Great idea! :) Thank you! :)
@@kingy-ai This would really be an amazing video. For reference, a self-help TH-camr called Hamza has really achieved this with his thumbnails. You can clearly see it's him in all of his thumbnails, so the person doing the prompts is able to use the same character doing different things based on the video's subject. I've been trying to achieve this with my own images but with no success. It would be amazing if you did a tutorial about this.
@@frankkosterschannel4416 I’m on it! V5 might be the way :)
@@kingy-ai Thank you man. I bought it specifically for this.
I have double checked and re-played the video several times to see if I made a mistake, I keep getting “invalid link” message from the bot
Still didnt find what i was looking for, you said you will go through till the abstract part, i do not see any abstracts. End up so much talking about the other artists.
I give up. I've been trying to do this for months with so many different pictures. It never works... 🙁
And when I copy the message link, and paste it into the bot, I don't get a second image - just the link spelled out. If I click on that, it just jumps up to the first image I dragged in.
Very strange. Just driving To my moms house. But will take a look when I get there. See if I can find the issue.
@@kingy-ai Thank you!
I'm annoyed. At myself of course. I have a personal photo of a friend, that I'm trying to turn into a more cartoony art style. No matter what artist I choose the image keeps coming out realistic. Help?
can you do this with urban landscape photos?
I don't know why but it keeps doing me the illustrator face instead of my media. Anyone has this problem?
Thank you for your work i'm a Newbie and this is helping a lot!
Bro is use stable diffusion 2.1 in Leonardo ai in custom model and i am facing a problem that my promt are generating bad eye and other pose with many problems😢😢
Ronaldo is indeed cartoon
Lol! :)
@@kingy-ai 🤣
@@kingy-ai btw. Are you intrested in some collaboration? That would help you grow your channel
Apparently can’t use image weight anymore? In version 4
trying to upload an image from my pc, the image goes away when i add /imagine. if I usse an image from online, it calls it "invalid" any suggestions.
so what if you got 2 images? how does that work?
My image doesn't have the option to right click link when in the feed?
Invalid link
The content type of the provided link is text/html, instead of an accepted image type.
i am getting this error message. how do i fix it?
You got a sub in me and you are my go to everything midjourney. I keep trying to use a video game screen shot and I want to either make it realistic looking or my comic book. All I get is imagine link is invalid. I upload it just like you say. I tried 4 different pics. Please help me what am I doing wrong? Do you have a video on upload game art? I am going to look. This was informative and I pray it helps
could someone pls help me? I drag images to midjouney and it wont let me add it.
Then I get "Invalid Link - could not validate this link". What sort of images do we need to use?
Hmm. Jpg and png usually work. Webp I’m not sure of. Sadly they haven’t fully updated the documentation for v5. But it should be out soon
NO matter what I do, it just makes a realistic image. Can't get it to make anything that looks like an illustration.
Can use this prompt in version 6?
I tried this several times and I can never get it to look like the artist in the original photo. I've tried --iw 0.5 and still it seems to combine the image with the illustrators model.
This is Midjourney..
You are not allowed to DM with slash commands
pls help..I copy the link enter in image prompt and do everything whatever is required ..I copy image link after opening in browser as by right clicking on the image it show copy message link with a link symbol a..I tried this copy message link as well bt after hitting enter the output is nowhere close to the reference image in fact it shows a different image only?? I have tried it multiple times but not working. Is it something to do with the settings?
@alinaminute thanx alot for this tutorial
I have questions please
Now if i generates ai art and i like it very much
So i want to replace the face in that ai art to my face
I tried alot but its always doesn't look like me at all🤦
Anyone have problems with styles library website?
Can you reference 2 images? For example one for the face and another for clothing?
Yes, you can use the remix command
How do you bring up the remix prompt?
i've never got good results (or seen anybody else get good results) using the 'remix' option...much better to just copy paste into a new prompt.
Hey there. Why cant i generate a original prompt today, even as a trial and with the /imagine command on, Midjourney tells constantly due to excess demand they cannot provide a trial version, to try in next day, or subscrive. What the hell?! Its AI! Thanks.
what if i don't want 'character by' and want a style? still use the same prompt or something different?
same question here :(
Hey man, thanks for the video. But the remix prompt window doesnt pop up when I click variations:(
Can you check to see if "remix setting" is on? You can access it by typing in /settings and hitting enter :)
I cant seem to create cartoon style images like you have with Ronaldo or Messi. How'd you go about doing this?
It is tricky - try this. character by neal adams::2, s.mj.run/JE0UsCsES7k --iw 1.0 --v 5
adjust the iw from 1.0 to 2.0 and the ::2 from 2 --> 1 if needed.
Ill give this a go. Thank you 😊
great tutorial, just tried it, used your prompt and got this failure message: cannot have custom image weights with --version4. Any idea?
switch up to V5! :) V4 doesn't support iw. :)
@@kingy-ai Thank you very much, did wotk, thanks and happy easter :-)
my midjourney is too V5, I did it your way but it's not like you, pls help me
How can I do this from my I phone
This must be a nightmare for all these unique and creative artists who are being blatantly ripped off and not receiving a penny in return.
doesn't work. putting girl's image your exact quote and getting a guy with beard (twice). Those yt midjourney tutorials are pointless.
What does it mean if it says invalid link?
Someone mentioned a Midjourney app, rather than using the Discord site? Or is it a Discord app?
Midjourney is definitely a discord app. That said, I just looked on the Apple App store and there is a free Midjourney "type" app - but it looks suspicious to me. It's called "Midjourney AI · Art Generator" on the app store. I haven't tried it as it it looks sus.
@@kingy-ai Cool, thanks!
I'm not seeing the remix prompt
type /settings and hit enter. Then click on the remix box (make sure it goes green).
I think I'm doing something wrong because I never get images that look anywhere near my pictures. I'm a man with a beard and I get women images, per example...
@Art Official Imagery Whats the picture for then? Taylor swift just imitated without a prompt
Same for me. I'm trying to change the image style or the place but trying to keep my identity, my face and I can´t get it. Let me know if you find the solution please!
after 20 attempts how can we create new account
those taylor swift faces dont look anything like her bro lmao
why do you sound as Andrew Huberman :D thanks anyway
I wish I had his brains and his muscles and alpha jawline. But it's coming :)
Yeah, so I've tried it with different artists and my photo, I'm not going to show the result to anyone, lol
Were they that bad??? 🧐😄
@@kingy-ai I set --iw 2.0 but the face was completely different. **sigh** I wanted to show a new cool technology to my kid :) I do enjoy using chat-gpt though.
@@frozen_eclipse Hmm... Did you try the ::5 or ::10 at the end of the input image? I don't know why, but this works sometimes.
@@kingy-ai thanks :)
Those artists don't get paid because of AI :(