Video

Wow, I have listened through several beginners guides on ComfyUI and I learned more from this than I done since I began using SD...

You're a legend!

Thanks for all your hard work and genius!

Thanks!

Now I just need a faster GPU... Excellent video - thank you.

LIke the others before me. A big thanks for those tuto. It's absolutly fantastic and useful.

What is the hotkey to add the reroutes to a node? Recording future videos with keyboard input capture displayed on screen would be super nice

I think for SD1.5 that works fine but SDXL has some issues with the IPAdapter, it gives a hazy effect. Changing the weight helps but when it's too low, the result doesn't transfer well, but when it's high, it gives that hazy feel. :(

You are the best! thank you very much. I'm getting error maybe you can please help? RuntimeError: mat1 and mat2 shapes cannot be multiplied (1232x768 and 1024x320)

Can't find Apply IPAdapter Node.

love you😭

is there a way to keep the head and hair but completely change the pose and the angle? Like for example if the character is in another pose from another angle?

This looks totally mind blowing! Thanks for sharing! Would love to watch a breakdown that is suited more for beginners, especially for the later part.

Amazing. Just to note that i get a lower Dist value after upscaling with NNLatentUpscale compared to not upscaling. I mean in the beginning of you tutorial. Even I used the same references like yours and same seed All same but not same Dist! Idk if that's normal or the quality of Pulid gets by Time better. But Amazing Matteo❤

Thank you for sharing!

First if all. What is ART?

I love you

For sure MagnificAI was born from this tutorial. Magnificent information.

Which node in maneger is “Face Warp", or github address? I can't find it.

Great video! Very informative ❤

Amazing!

Why doesn’t it work so well with the SDXL model? In my case, the result is one girl with different styles on two sides of the head.

I'm looking forward to experiment with this! One thing that stands out to me out of the gate though. Shouldn't I get the same render if I pass the same prompt to a regular sampler and to the all pin on the injector while using the same seed for both? What is the technical reason for getting two slightly different images rather than two exact duplicates?

is there any demos about advance ipadapter features？

Thank you so much for your lessons, I made a short AI film using your lessons, I set up IPADAPTER and AnimateDiff and got an amazing result, I like it very much czcams.com/video/MFmTsgkLXRA/video.html

the latest IPAdapter node doesn't have a noise input. How do you set the noise in this case?

Why would you want to daosy chain two ipAdapters and two ksamplers in your initial unified loader workflow? Does that help with Fidelity and consistency?

What does the ConditioningZeroOut do?

do we have tile controlnet for sdxl that may work with comfy?

Valeu!

Bring the simplest node as you said in the last of the video

It was a great education and a great inspiration and lesson.

Very interesting, thanks for the great work!

Thank you for all the hard work

You are the real master and you are so generous to share those techniques. Thank you so much.

This is something I've been thinking and talking about. Your video was very insightful, and helped me put into words some things I was struggling to describe. Thanks!

Nice work! It doesn't support embeddings :P

Hi great stuff, is there anyway to do this kind of attention masking with loras, so I can apply separate loras to separate masks? There's a few things kicking around but nothing seems to work all that well.

Well, it is not easy to find someone you totally agree with everything they say, and you're not an exception. Don't sell yourself short. Your dedication and efforts make you one of the most important individuals in the world of AI. That was my opinion, and today, with your insight and talk about ethics and controversies of the use of AI and its future, I am more convinced.

Matteo's creation of such precise tools elevates technology to the level of art. The specificity inherent in these tools rivals that of art itself

Hey Matteo, thanks for the amazing job you're doing. Following this workflow i get an error: "only integer tensors of a single element can be converted to an index". This is happening when i turn the IPAdapter Batch nodes "weight" widgets to inputs and connect them to the IPAdapter Weights node output. Somehow if i turn back those weights inputs into widgets, the Sampler is able to process them, but ofc i don't get the desired result. Do you know what this might be related to?

i dont have sgm_uniform as a scheduler. can someone point out how/what to get this?

Excellent tools, thanks for sharing

When Matteo says suffer, I just think joy! 😄

8:47 that does NOT look alike Korean at all

A beginners guide course on comfy ui with everything you know would be amazing!!!!...................most of your videos look like they are geared towards intermediates and experts

What if the role of each UNet block is different for different unrelated models (like SDXL vs Pony - they are seemingly the same in sense of architecture, but merge of SDXL and Pony gives bad results).

Great work as always Matteo. How does PuLID compare to ipAdapter in terms of vram requirements? I can run most of your ipAdapater workflows even on my puny 4gb VRAM, but I imagine PuLID has higher requirements.

Still experimenting. This what I found....Checkpoint Model matters....otherwise its still a bit of a roll of the dice with hentai type models.....even using your direction as per video. In one prompt (2 batches), I asked specifically for a girl with dark brown skin and got....a pic of a curry meal and the other image what I prompted for....LOL . I'm realising there are limitations to the number of words you can use, so no fancy or detailed promoting I'm afraid..... Others I'm getting random images of animals and sometimes classical pictures?? The style prompt in particular (output1) has a big influence on the prompt outcome, need to be very specific with language around style (interesting using words 'digital art') . Think I will wait for further development before incorporating into my other workflows. Thanks for sharing anyhoo