Видео

Love pixelwave too! I found it's particulary good for nature photographs.

I have been looking for this and couldnt find it and took photoshop with generative fill, well this is better and free ! it's just missing an inpaint function to add extra detail and sometime the light is weird. it's definitivly close to be perfect now i just need to get a better graphic card 🤣

send me prompt

This is easily the best channel for people trying to run Comfyui with limited VRAM. The info here is nothing short of amazing. Thank you sir

I was wondering if you know of a Flux 1.1 dev node for Loras (16 b)? (comfy UI)

ComfyUI makes my head hurt! But, every time I see a video like this i have to try again. {sigh} God help me...

Do you use the left over noise option? In the kssmpler advance if you pick the left over noise it would send de image without denoising to the next ksampler. Also try without send left over noise and inject noise after each ksampler separately with a noise node. it add more details those 2 options should improve the 3 ksampler setup and give each ksampler the option to fix errors in the composition

Most amazing workflow I have seen! Excited to see the new workflow when ic-light is compatible with flux!

Would be awesome if you could further optimise this technique with say Impact nodes cfgschedulerhookprovider and stepsschedulerHookProvider nodes. I havent checked though if they work with flux, but they should

excelent, works fine, but so slow..

Thx man, useful vid!

Hardest installer Custom node for me...I cant install

Thanks for sharing this!

Thank you so much. Stunning quality with Pixelwave 🤩

Hi I am still figuring out how to put exact clothes on the Modell or even have a product in the picture. Only way seems outpainting to keep logos and text correct. Not even catvton is able to keep a design of a t-hirt :( Basically what I need is a a combined outpainting and inpaintng workflow. One picture to fit exactly into another, without new prompting, if you know what I mean. Is it possible?

Fantastic work....in my case it would be great to have this workflow in a Flow style. Did you ever think about it? thanks

Which laptop do you have and how much VRAM? How long does it take to generate an image?

@xclbrxtra Hai brother, Your videos are 🔥. I want to know that should we really want to terminate the pod after we use, will it chargee for not terminating??? I dont want to terminate my pod .So, that I can access it anytime. I need your Suggestion

how to avoid character getting cropped?

Only issue with Flux is the fact that these trained models like Pixelwave won't like LoRAs that much, eyes and smaller details get distorted. Maybe some folks will eventually train on top of those de-distilled models.

gguf again? any workflow with dev original or dev fp8?

why gguf?....i cant bilieve, most workflows i see interesting are using gguf. I use dev original....dosent work. any advide? thanks

My dumbass thought that was a pic of Riley Reid lolllll. Thank you for the video. I actually decided to use ComfyUI for the first time and it took an hour to install everything but this workflow works. Keep up the good work!

pixelwave is awesome

Every other day a new Midjourney killer! 😀😀😀

train a lora, done.

Thanks Bro for this tutorial

thanks, how about load an image , can PixelWave can handle this to transfer load image to another style , and even further with control net ?

Thanks a lot, this is very helpful

--tile or seamless feature is still not available on flux side

bro, what comfyui version you use, also the python and pytorch version pls

Isn't Flux already considered to better than Midjourney ? I mean since the release of flux, all I've been seeing and hearing is how it's a MJ killer and whatnot.

Как всегда ты супе чувак. Спасибо за воркфлоу удачи тебе

Great small website, but the price you described are wrong... Minimax give you 100 per days, misconception by a lots of people, that's 3000 extra per month for a total of 4k and + your free 3 days that's over 5k+ =200+ generation for 10$ sadly sellerpic is still too expensive.

Hey man, great video. Can you give me a hint on how to do this using a different image? I have a blazer in another photo, so that image would need to be the input. Can you do that and show me?

This works! Thank you for all the extra explanations!

how can we do this with a batch input with thousands of images that need fixing?

Does this only work with the portable version? I have downloaded more than 20 workflows, I have watched more than 30 videos and none of them have worked for me, this is one more, thousands of different errors without a solution, I am already losing hope of making comfyui work

NO FREE TRIAL, AFTER CLICKING GENERATE (10 CREDITS) there is a new window where it push u to buy plan anyway

I got issue with Math Expression. It gave me this error, can you help me check? Prompt outputs failed validation: Return type mismatch between linked nodes: a, FLOAT != INT,FLOAT,IMAGE,LATENT MathExpression|pysssss: - Return type mismatch between linked nodes: a, FLOAT != INT,FLOAT,IMAGE,LATENT

I did everything as advised but keep on getting the " UnetLoaderGGUF too many values to unpack (expected 2) " error. running on a 3060

thank you for this very interesting node MAGICAL but i don' t understand whiy the switch image1/image 2 with the node rghtree doesn't work, no reaction when i move the cursor on the area, what can i do ?

man you are a legend, i have followed many tutorials and wasted many hours, your video is the best, easy to follow guide and the correct links provided was super helpful, thank you. and yes everything worked great, finally!!!

thank you very much for this good job but can you facilite me to indicate chere to find your workflow for this video? i don't find it anywhere (sorry for my english); if you prefer just said to me the label of previous video where i can find it.

Hi, i have a trouble to query your workflow, because i got error with "`newbyteorder` was removed from the ndarray class in NumPy 2.0. Use `arr.view(arr.dtype.newbyteorder(order))` instead. " .. do you have any idea why? Are you using Python 2 or 3 ?

I'm downloading flux1-dev-Q5_K_S.gguf, while I watch this. Hope it will work on my 3060 /w 12gb

I'm sorry TripleCLIPLoader is Ambiguous as is UnetLoadGGUF also. I have ComyUI v0.2.5-3-g65a8659. These nodes will not load for some reason (But why). Why are we defeated by Ambiguity in this age of AI. I'm speaking of Computer Abstraction. We must bring along our people, we must Excel? This is not a TROL. But we need to remove Ambiguity in our communications if we are to progress! This is a great video, however, Nodes are ambitious. If we continue to struggle with simple configurations while we are otherwise experts, and in the age of AI, we are defeated buy our inability to communicate without Ambiguity. I'm sorry I'm a classical Computer Scientist with limited time in a Life dedication to this new application. For every positive there are 10 negatives who will not express because they don't have the time, but would like and need to progress. I guess we must live with we cannot rise above. I would give this a grade of 'A'. However. it is Ambiguous. Ambiguous in some areas of communication, However, generously given which is a tremendous sign of dedication. Thank you! But you have "More To Master". I mean, Come on, where do you put the specific files of TripleCLIPLoader and UnetLoadGGU for my newest version? We need to get it right. If I write a Kernel Driver for Linux. It needs to be explicitly correct or there will be bugs. Which is not acceptable. This is simply, A critique to do better. Regardless of the software. You getting close, but it only works for a few people.

Great video so there is no need to stop using Flux for SD 3.5.

as always respect you, sir. have a nice weekend, sir.

is this IC-Light V.02? :) better