Consistent Characters Soon | Midjourney Office Hours Recap Feb 28th 2024 | Midjourney News
Вставка
- Опубліковано 28 лют 2024
- Exclusive Prompts ➡️ ftp.live/patreon
Prompt used: illustrations of [subject] drawings and sketches, in the style of collage-like compositions, complimentary pastel colors , abstract portraits, scanner photography, eccentric detail placements, gestural markings, vibrant illustrations --v 6 --s 40,60,400,800 --c 2, 6
- Thank you to alexoo8521!
Podcast Link ➡️ / @futuretechpodcast
► FREE PDFS
Prompt Glossary V6 ➡️ ftp.live/V6Glossary
2-4-2 V6 ➡️ ftp.live/242
200 Aesthetics ➡️ ftp.live/200Styles
Midjourney Cheat Sheet ➡️ ftp.live/CheatSheet
► SOCIAL
Discord - / discord
Instagram: / futuretechpilot
Thanks for the update. Sounds exciting. Great looking images with that prompt, BTW.
haha I definitely seem to favor these kind of images, I'm glad other people like them too!
new version style reference seems interesting
Next week already. Very interesting!
The progress always blows my mind
Oh my gosh i really want to create a comic. It haunts me how much I am so bad at drawing. I’m glad I live in this era, I can finally make my childhood dream come to life. One thing I’m worried about is how they’re gonna pull off dynamic posing, for example action scenes. I’m sure they’re gonna do it in the near future. I’m literally dying to see what midjourney would become.
I completely agree! I have no visual artistic skill of my own and Midjourney is a dream come true
I made it on time for office hours this time. It was so fun with everyone posting pictures!
I didn't stick around for the part with Grimes but I'm glad to hear it was fun!
Wow ... really surging forward.
Thanks for posting these, really apprecieated.
And I always appreciate you taking the time to comment!
Thanks for the update man ! Much appreciated!!
Cheers pal, thanks for tuning in!
Cheers for doing these weekly updates, really appreciate it!
And thank you for taking the time to comment!
Thank you always for the info! So excited for what's in store next for MidJ. Keep on being your awesome self!!
And thanks a lot for taking the time to comment! The future should be fun!
I love this prompt!Thank you
Cheers!
You could have an art gallery opening w all of these amazing unique style images!
😂 maybe one day
Thanks again for the recap! Random style sounds exiting for me, I love to explore styles and experiment....
I know right? It's so much fun to see what's out there
Thanks for providing the office hours recap!!!
Cheers pal!
I don’t understand how David can be so excited for V7 when V6 needs a lot of work? If they’re struggling with V6 on certain things why is V7 so promising?
they’re going to start deploying new versions faster instead of fixing old ones which is the right move to make instead of taking 8 months to release a new version like they did from 5 to 6
@@KingZero69 Isn’t that bad to rush things? I’d rather them take time. If anything, version 7 will just be V6.1. I don’t see how the difference would be a huge difference.
@@dwayne6381 i get that take that they’re “rushing” but i think it’s more that they’re “retraining” a new model for each version instead of just trying to tweak an existing model that has flaws
It's a good question, but I think it's because of how fast this tech changes and how much the team learns by experimenting. These models take weeks to train, and Version 6 was made like three months ago - so they've learned a lot since then. And they can't just put that new knowledge into the current models, you have to retrain them - which is why Version 7 will be better.
@@FutureTechPilot Thank you man. That makes sense.
Midjourney with Chat GPT would be really interesting to see. I use both, but have defaulted back to DALLE for the last few weeks for my story creations since the ability to create your own GPTs came out - which helps maintain overall style, scenes, and characters. If they're able to apply common language prompts to Midjourney through GPTs, that would help people like me who have trouble getting the images I want due to, at times, poor prompting - as you went over in your previous video. On the flip side, if this were to happen, you wouldn't have the creative freedom to mess around with copyrighted content due to DALLE's policies.
Ohh that's interesting. I never thought about Midjourney being even more restricted on other platforms. fascinating
Can’t wait!
💯
This is why I subscribed, for videos like this. I was blown away by the images you generated; they are what I try to achieve in Midjourney. This is also why I love MJ. Not one other AI generator I've tried puts out images with this kind of sophistication. Plus you are a fellow Canadian, so there's that. Thanks for the vid, you're doing good work.
haha I know what you mean, MJ is the best! And thank you very much for the continued support
Hi, two questions: 1. Is there a referral system for when we encourage people to sign up for MJ, we get something for it? 2. What’s the deal with the artwork I generate often has a faux signature in the bottom corner like someone drew/painted it? How do I keep this from happening? It is extra time for me to have to crop/edit the faux signature out of the corner of the artworks and it seems to be happening more frequently lately but not 100% of the time.
1. no affiliate program, they don't want their users to become sales people
2. you can try adding '--no signature, watermark' to the end of your prompts, that might work!
@@FutureTechPilotThank you
I don't understand, will they release the character consistency option next week or is it a hope?
Hope. They're testing it now, if it doesn't meet their standards, it will take longer
I like the videos! Thanks for making them.
One piece of constructive criticism; the inflection in your voice tends to go higher as you finish a lot of your sentences. I think if you try to vary your pitch and delivery of various sentences could be a good change. Hope that helps.
I really appreciate the feedback!
I'll stay cognizant of that.
I'd love to get access to an API myself and it was available for developers generally. I'd hope if they were doing limited partnerships, it would lead to an open (paid of course) API in the future.
Yah I wonder where we will be in a year!
I want a image of myself with different graphic designs for my music singles
It sounds like you'll be able to do that soon!
@@FutureTechPilot cool 😎
How come smaller establishments like Scenario AI using existing models can already give you consistent characters but larger companies like Midjourney are still struggling?
Open source models contain progress from the entire world, while Midjourney has their own private model. The downside is that it's difficult to create features (they have to do their own research and experiment) - the upside is that their model is the best
@@FutureTechPilot I agree their model is the best but so far, the other ones even with limitation have been more usable to creators and artists with serious and continuous projects than Midjourney. With the recent sref though it's good for generating consistent style images to train your own style in SD.
I subbed for the year rate a couple of months ago. Will there still be a benefit for the poeple hanging out with some room keys they paid for? ;-)
😂
Hmmmm, just after ive set all my --sref's into presets......
😂 I feel your pain
Hello can I also give them a feedback... possible to train the model to do Chinese calligraphy in the future? right now I don't think the model recognise those characters... thanks again for all the info you shared!
You should suggest that in the new feedback feature! type /feedback into the prompt box, and you'll be able to post your ideas there
uh didn't know there's such function... thanks much again! @@FutureTechPilot
hi, maby you know if their AI is connected to the network or they constantly add data separately. why for example, midjourney knows how looks logo of Superman, Batman, Spiderman , but does't know how looks other famous logos/symbols. for example, the Witcher or the Harry Potter logo with 4 faculties, or knows famous characters from one anime, for example Naruto but can't generate another famous character Satoru Gojo (anime Jujutsu Kaisen) normally. What I mean is that he knows some famous themes/characters/logos and doesn’t know others. why is that ? another question ,do devs plan to add a 3D ilusion pictures prompt option?
If Midjourney can't create something specific, it just means that those logos weren't in the dataset used to train the model! and there's nothing we can really do about that.
As far as 3d illusions, I love that idea and I've tried that prompt in the past! haha but I have no clue if they're ever going to make it possible to create properly
@@FutureTechPilot ok, got it thanks for answer.
Hey bud, do you do professional consultation?
I will soon! Working hard to put out my Beginner Course and then I'll set something up for 1-on-1 calls. Thanks for asking!
@@FutureTechPilot Eagerly waiting for it. I use MJ to create graphic novels for a long time now but there are a few areas I feel that can be streamlined so that my output can get better and faster. Drop me a message whenever you start with this. Cheers
You know, I used to really hate you (true story), but now I like you! Also thank you for spelling "complementary" correctly. Well done! That is a compliment.
😂 cheers pal
👋
🤝
if they can pull off creating a realistic and consistent human face in different settings and styles that will be quite the magic trick. sounds like its still sticky for them in that area but it sounds like they have it down for cartoon characters though lmao
hahaha yah at least a cartoon character
Are they jumping in with Elon???
I have no idea. My guess is GPT
If MJ partners with toXic then all of David's talk about bringing goodness to the world is a big pile of bull.
My guess is GPT
I am not so whelmed. So consistent characters sound VERY general. And only "nice images"? Seriously?
Go to a gallery showing the baroque art of Caravaggio and then tell me again why art needs to be "nice". What a rubbish idea.
We'll have to wait and see what it's like at launch!