How to use Midjourney Image to Image (Best Image Prompting Methods!)

Рет қаралды 22,679

Күн бұрын

Пікірлер: 39

@morpheus2573 Жыл бұрын

Thanks for sharing those image prompt tips. Not sure if you're aware, but a few days ago, MJ added another feature to image prompting without really announcing it. They've added the functionality that allows you to weight images individually. Before, if you stacked a group of image URL's in a prompt you could only weight them as a group, ie --iw 2 Now you can weight each image individually, in the same way as word weighting. ALSO, it works with permutations. This allows you to experiment with various weight combinations within a single prompt. eg... Here is the same three images in two separate groups of three images, each with completely different weights. The two groups are separated with a comma, using curly brackets to create 2 x different permutations within the same prompt. {IMAGEPROMPTURL1 ::5\, IMAGEPROMPTURL2 ::2.5\, IMAGEPROMPTURL3 ::2.5, IMAGEPROMPTURL1 ::3\, IMAGEPROMPTURL2 ::5\, IMAGEPROMPTURL3 ::2} --ar 16:9 --style raw

@WadeMcMaster Жыл бұрын

Lol typical, I find out right after I publish a video 😂 But seriously, thank you for letting me know!

@WadeMcMaster Жыл бұрын

I pinned your comment so hopefully people see.it

@morpheus2573 Жыл бұрын

@@WadeMcMaster IKR? Innovation can be so inconsiderate. 😉

@morpheus2573 Жыл бұрын

@@WadeMcMaster 😎👍

@leogiron02 Жыл бұрын

can you explain a bit further? brand new to this

@caligaselcabron Жыл бұрын

Yeah your content deserves more views, this was super helpful. As always congratulations Wade, wish you the best buddy!

@WadeMcMaster Жыл бұрын

I greatly appreciate the compliment, thank you!

@wendyon4517 Жыл бұрын

This was very helpful. I've been having a heck of a time repeating a specific design with a dog. It only works with one breed I've prompted. This short video solved it... I think. It worked with a similar design 😊.

@WadeMcMaster Жыл бұрын

Glad it helped! Thank you so much :)

@tomkelly4351 11 ай бұрын

Really useful stuff! Helped me a lot, some good tips! The only thing I might say is the images you used are very busy and of a particular style. If you used more general images and images in which you can really see the blend happening, like contrasting images, I think it would get the point across to a larger audience.

@WadeMcMaster 11 ай бұрын

Good point! I do agree. Thank you :)

@HellNahAI Жыл бұрын

This is a big help! I upload all of my creations onto KZbin. Advice is welcome 🙏🏾 😎

@WadeMcMaster Жыл бұрын

Man, keep doing what you're doing, I love the style. Especially like the waking nightmare - some cool stuff!

@flatlineracer3207 Жыл бұрын

Your Amazing Wade. A wealth of knowledge. Thanks for sharing!

@WadeMcMaster Жыл бұрын

Thank you! Much apprecatied :)

@oligei666 Жыл бұрын

Thanks for the information! Its really useful !

@WadeMcMaster Жыл бұрын

Glad I could help! Thank you :)

@tinas6459 Жыл бұрын

Wow. This is amazing! Thank you

@WadeMcMaster Жыл бұрын

Thank you!

@puja1985 Жыл бұрын

How does this image-to-image function compare to Leonardo AI's? Which is more user-friendly and effective?

@WadeMcMaster Жыл бұрын

They both do interpret things a little differently. I think Leonardo Ai's is more user friendly but I've seen mixed results on both - some better on leo, some on Midjourney. I think Leonardo's is better at referencing the image but I'd need to thoroughly test it.

@historiasdelodesconocido5021 11 ай бұрын

Excellent tutorial!! a query I have tried to upload my own design of a character that I have created in a unique position, but the problem is that when I upload my design by writing to midjourney to add more characteristics to the body of my character without changing the position in which the character It is found, it shows me different poses and not even my design is the same, much less in the position it is in, can that be achieved with this AI?

@WadeMcMaster 11 ай бұрын

Yeah sort of... If you made the original image in Midjourney, I'd add thos pieces on with Vary Region - or you may need to upload the image to something like Leonardo AI's Canvas - or use Generative fill in Photoshop to make those changes in specific areas.