Відео

If I want to liquefy a certain part of the human body, such as face slimming and high cheekbones, can this node be achieved? Are there any nodes that can liquefy a certain part like photoshop does?

I'm watching your videos and they are great! Things just work as expected and I have to say that is amazing (I watched a lot of random videos that look amazing and after trying the workflow the results were bad). You know how I can use an image as reference to other for the inpaint? I'm trying to generate multiple images from the same person but I'm having trouble with the clothes or other details that change all the time (like diferent figures at the shirt, diferent cuts at the jeans, different earings...). I will keep searching for the best solution but it would be greath to have your opinion about it. Thanks for the videos!

This works whit anime style?

i don't have the queue promt windo, have no idea what i did wrong

Very high quality video

I am interested in one question how to put the nodes in straight lines in the same way

Great set of videos- thanks for posting! Heading over to your Patreon

Hi. I am a member of your patreon. I just tested your Aris Workflow 7.json I get an error with the LDSR Upscale (all-in-one) node : LDSRUpscaler 'state_dict'. Please; Where can I send you the complete error?

I believe that the way in which you have presented this information will only perpeturate confusion surrounding the Flux Dev License. You may not use the Model or it's derivatives (finetunes) commercially such as an inference or image generation service. For example, Civitai may not make available flux for use by it's users without paying a license to BFL. However, it states clearly in the license: "Outputs. We claim no ownership rights in and to the Outputs. You are solely responsible for the Outputs you generate and their subsequent uses in accordance with this License. You may use Output for any purpose (including for commercial purposes), except as expressly prohibited herein. You may not use the Output to train, fine-tune or distill a model that is competitive with the FLUX.1 [dev] Model." This means you can use the model to create images, and sell those images. You may use your generated images, the output of the model, in any way you wish, including commercially. The exception is you cannot use the output to train another model "that is competitive with flux". The portion that specifies "except as expressly prohibited" is in regards to the list of things you're not allowed to make, like illegal images.

Subscribed... You seem to know your way around comfy too... BTW can you code, python? NodeJS? JSX maybe? React? Cos I'm on the lookout for folks, any folks, that can code, know comfy, understand inference and how LLM's work... and are looking to join a team building a commercial app I'm building... Just saying.

I dont have civitui, i cant be the only one cmon guys

Great video … I am at the point of creating a character sheet with controlnet nodes for Lora training. Hope you video does loa training inside comfyui and not with Lora gym or whatever it’s called.

what a great tutorial super details everthing

Great video, got it working in minutes by following this guide! I'd love to see an example of adding a LORA to this workflow? :)

To fast very confusing...

I don't know what is wrong with my setup in comfyui but on a RTX 3060 12 GB vram, 64 GB ram with GGUF Q6 model with one lora and the image resolution is 768X512 and it take 15:40 at 20 steps.

Thak you for the vidéo. When I opened ComfyUI, I don't have the window on the right (Queue Prompt, Save, Load, Refresh, ..Clear..)!

Not sure if there is a considerable improvement gained of the Image quality …

The comparison is super helpful

seems like a great tutorial, but you go so fast and there are actually no links in the description for these things. I did find the links on another page through a link, but good god man, slow the hell down and zoom in on important info such as file names and dont quickly go past important information such as unzip file or drag file. The snapshot is a long folder name, and I did not see that folder name, so do I rename it or just leave the folder name as is or is there a folder inside the long name I need to transfer. sheesh, if you spent 3 seconds instead of half a second on the drag folder page, I may have been able to follow. The tutorial is amazing but I am having to pause, reqind, replay way too often

SDXL is stable?

Good Video👌

LoRA Training NEXT!

nice simple straightforward explanation without deviating or meandering on side issues. was looking for a simple vid that explained how to set up a simple facedetailer.....but most are 30 to 40 minute vids with a surrounding expert wire sphagetti workflow that leaves me cross-eyed. thanx.

sir please make a work flow for constant character with prompt + face + pose + cloth + pose + additional products please , please , please ☺

I had some more unruly hair and it didnt work very well. The model was generating faces and weird stuff where the hair was supposed to be. Think I'll do it with photoshop for now :(

My video suggestion is the use of other Non-image models such as for example LLM's, or 3d generative models and even directed movement "move here, do this, do that" sort of thing

keep going , sir. thank you.

Thank you for the tutorial! And do you know what the difference is between Xlabs controlnet vs controlnet union by InstantX? I mean why did you choose this Xlabs version, and not the controlNet Union by InstantX with more than 3 options?

very helpful Thanks

Thank you so much for watching! I’d love to hear your feedback-if you have any tips or suggestions on how I can improve my videos, please share them with me. Your input means a lot!

This is amazing

I notice my generations stop short. Is there something that I need to tweak? I'm running a 3090.

Very interesting 👍

2:04 definitely picked the right spot to select, lol😂

I hope you enjoyed the video! Feel free to leave feedback so I know what I can improve. The next video will be about ComfyUI Loras - so stay tuned!

The video was helpful thank you.

I'm back with a new video, and from now on, I'll try to upload every 5 days. If you still have any questions that weren't answered in this video, feel free to message me on Discord.

bat file did not run.

Can you help? Pressed 'open in sam detector' multiple times but nothing come up.

I just spent 2 hours installing and uninstalling this just to get error messages both times. I tried to join your discord, but I'm not giving you my phone number. you should answer the error questions here under your tutorial

Failed to load the extension "elevenlabs_tts".

Thanks for the tutorial! How can I download an image and extract the workflow from it? A regular download doesn't work

The best ComfyUI explanation on the internet, Thank you so much for this

can you make 2 girl friends on the same chat so they can get jealous of each other?

This is fantastic!

Is it possible to do it on video too?

can you plug this differential diffusion node into any impainting solution? Combined with COntrolNet for example. Edit: yes it does. Just plug in the Differential Diffusion node and you're good to go. Just that little node changes everything. What a discovery, wow

Underrated! You deserve more.

2:28 if you're using like an RTX gpu it shouldn't strain it for recording since it should use the Tensor Cores for the AI generation which wouldn't affect the processes used for encoding video