Видео

Hey I appreciate that you’ve done Mac comfyui tutorials- I’m running a Mac Studio ultra m2 with 128gram…but can’t figure how to get it to use more than 70% of the ram and speed up processing. Any advice is welcome.

Thanks, really liked this node. It's simple, runs really fast and has a good result. Great if you need small variations of a face.

Thank you so much, i really appreciate your tutorial

Very helpful thank you!

Can you do a review of mimicPC? Curious how it stacks up to ThinkDiffusion (which i love)

how can we combine this with crop and stitch

the only good tutorial on youtube

What's with the weird bait and swap? Goin by thumbnail, I thought I was clicking on something that wasn't a tightly/no clothed female again, but, nope!

Thanks this was helpful. Appreciated the step by step visual through the detailed steps while being to the point.

you lost me at the torch install. it's just not working for me :(

Great tutorial. The only hiccup I experienced was I followed your tutorial, as well as the Apple-specific instructions GitHub, which had me make the ComfyUI folder within the models folder. This created two ComfyUI folders, which was incorrect, or at least was confusing in regards to which directory to have Terminal point when installing the requirements.txt file. Ultimately, I was able to figure it out. Do you have another video about installing models-which models are used for what? Thanks!

Great stuff, we do like 1111 over comfy, but one thing: order is also important. that is to say. if you create a generation and aviator glasses is at the bottom it might not show, move it to from of your prompt and the will return.

If it helps anyone that's still getting average results even after the fix mentioned in the video, using dpmpp_sde instead of dpmpp_2m for both Ksampler and FaceDetailer gives me perfect results on 'full body' type images where the face is incorrect. KSampler steps are 40 with a CFG of 7. FaceDetailer 50 steps with a CFG of 8, with 1024 guide & max size. Scheduler is Karras. You could probably reduce guide & max to generate quicker, haven't tried myself yet but, good luck!

No art, no skill, just laziness. This is dumb.

Great, Definetely worth a try

How much minimum VRAM do I need? 4GB? I have a GTX 1650

OMG thank you fo this simple tutorial video! I just getting into ComfyUI and only have like 1 month using webui forge previously. ComfyUI seems overwhelming but your tutorial here is easy to follow and very on point! Thank you soo much once again.

Hi, I installed ComfyUI and installed the checkpoint. But when I run the queue, there's a popup said that "Error occurred when executing IPAdapterModelLoader: Weights only load failed. Re-running torch.load with weights_only set to False will likely succeed, but it can result in arbitrary code execution.Do it only if you get the file from a trusted source. WeightsUnpickler error: Unsupported operand 60" Ive tried updating comfy, downloaded the nodes directly from repository and so on. Cant find any info on this error." Hw do I resolve this? MY checkpoint is "v1-5-pruned-emaonly.ckpt" that I downloaded from hugging face

Love those quick tutorials! I've been using IPAdapter to try making consistent characters but the results aren't always great so this will definetly help! Would love a video on this if it fits into what you’re working on!

how can i add the empty latent image to change the size and the batch?

Check out ThinkDiffusion here: www.thinkdiffusion.com/?via=promptingpixels

thank you very much! i become stupid with deforum trying to achieve that kind of result!

I got stuck on adding checkpoint file

In today's RUclips, it's rare to see someone like you who gets straight to the point. You also teach well - calm and clear. You're golden, man.

I really appreciate this video. Learning how to use comfyUI can be a daunting task and I like how you actually explained what was going on a little bit. Copying a workflow can be nice, but this video helps me to understand a little bit more of what is going on. That's very helpful because it allows me to take more control of my generations and be able to tinker with the settings and get more specific results. So, thank you!!!

bro is running from the police

Thanks for this. Do you have experience with using external hosting services? I want to transfer a workflow, models, and various nodes to a server and have the workflow available via a simple GUI of it on my personal website.

does it matter if the workflow its not loaded to comfyui? let say I want to use multiple workflows together?

On the local install of ComfyUI portable, it no loger shows the buttons like Save, Load, Refresh, etc. It only shows the queue button. Thanks for sharing the keyboard shortcuts. So others know, they moved the workflow GUI options into the top navbar to save screenspace with queue button. Loading, Saving, and Exporting worklows is under the "Workflow" header. Undo, Redo, and Clear workflow options are under the "Edit" header.

Could you tell me how I can, instead of sending an image, send a video to be processed by the reactor?

thank you for this class, it really helped me a lot to understand how to automate using python

I mean, if you were wearing a white hood, I'd likely be more dismissive of what you had to say.

Are you AI?

Personally now I second guess everything I see on the internet. I saw the photos published from the Indian space agency with photos of the lunar lander and, I couldn't convince myself I could trust the photo I was looking at anymore because of the prevalence of AI.

thanks a lot. i really need this. i want to know, do we node for sending audio and receiving video? in python.

Thank you good video, only small request I wish you would say what model you use...

While this method works for calling ComfyUI APIs in Python, the WebSocket API calls are pretty unreliable. They keep disconnecting all the time.

thanks this is great!

Thank you! Spent hours banging my head against the wall. I was using "ClipLoader (GGUF)" I needed "DualClipLoader (GGUF)" because I was apparently missing "clip_I.safetensors" which results in errors. This video showed me which one I needed to use and it works fine now.

Ff

I want to try something like, using segmentation for mask and using an image to replace the mask, so I want a character holding or showcasing a product, either by wearing it, the product image should be passed as image input to replace the mask, is that possible?

you need to say go ahead one more time

2:32 i dont get it. What are you doing there? where does this folder and the terminal come from????????

I literally don't even know how to SAVE my darn workflows. I can't find the button anywhere. The buttons you are using do not exist for me.

SO I installed stability diffusion with both comfy and automatic1111, The "Image CFG Scale missing on img2img" tab Checked my version and even tried installing the depreciated pix2pix, still no image CFG scale on the img2img tab.

Newbie question but what's your shortcut to search across installed nodes when you create a new node?...

What node should we add if we want our own custom output image size? I have “empty latent image” node, but there’s seemingly nowhere to attach it with the “load image” node attached.

Thanks my friend, I am not sure if without your video I would have managed. 🙏

We need a one click installer...

Thank you for the video. The topic is completely new to me and I would like to make AI Video Avatar for my website with my face, multilingual etc.. Would I be on the right track with this tool? Thank you for help and answer.