Відео

These steps will alienate a significant number of people. Hopefully, they can do some kind of comfyui approach.

weird it says Lora but it dosn't have the Lora tab?

stays in, launching roop, and nothing happens

hello, if i want to add a node , for exemple comfyui-reactor-node is possible?

incredible job. extremely clear tutorial

does it use xformers in this version?

Thanks a lot! Works on MacBook m1 pro. I'm new to ai image generation. But after watching different tutorials I must say: what a nice and very helpful community ❤

Thank you for this. I don't have the manager button at the top right. How do I get that please?

i rarely comment but i learned a lot and you helped me decide which one to use . thank you for taking the time . i'll be sure to check all your videos out later.

It works great! and working ComfyUI node plz?

fantastic video thank you

Can I install on Windows 10?

i use mac intel 2020, i cant install.

Excellent instructions. I was able to install everything and Control clicking does open Omnigen. However, with every image I use as well as the sample images and prompts on the Omnigen page, I get the following error message: TypeError: cannot unpack non-iterable NoneType object. How can I fix this?

In the generate base face image, i think the inpaint looked the best. it may just be my eyes

Best tutorial that actually works. Thank you so much. Keep making such content

My machine cannot take Nvidia GPUs, what do I do?

Hi Foxton, what computer minimum spec to run and install this liveportrait ?

Overjoyed to have found someone who not only knows his stuff but is also a good teacher. In computer land these qualities rarely go together. The sound quality is good, the speech is clear and slow and steady rather than garbled, the flow is concise and clear with each step explained before the next one. Unusually. instead of expecting you to follow a cursor all over a page at impossible speed, there are excellent diagrams and screenshots which are held long enough for you to hit the pause button. Classic teaching skills!

Great job very helpful to determine trade offs you make by using the diff models.

Oi, thanks mate! I had errors when putting it in a folder of my choice; but when I moved it to the root directory of hard drive, it works fine. 👍

Hello! Perfect and easy to understand, I have one problem at the end when creating a textfile, the window "Do you whant to change the fileextension" never shows up so when I copy and save the "Syntax 6"in notepad it's still a textfile... Bengt

Why why I did exactly why you did but I can’t find it in the checkpoint I need help pls

Nice video!! Does it work in a PC with Windows 7 (32bits) or at least in a notebook with Windows 11?

I keep hearing Q6 is the "sweet spot" but no one seems to have comparisons with it.

I wouldn't install any of these faceswap fantasy fulfillers.. it seems like the latest viral scam for hackers to get users to install their malware.

How do you set the interface to dark mode, I am looking for a --theme dark sort of setting but can't find anything, and trying the settings within the browser GUI I cant find a dark mode but I can find dracula etc.

I have A1111, ComfyUI, ForgeUI the amount of dependencies that cause so many problems is unbelievable. I am wondering how this one will work.

Hello - Great video finally I managed to change face with video and image. Other tutorials where not good enough. What`s the best face swap tool for image ? I noticed when I`m changing picture where girl has a hat on her head it`s affecting the hat itself - face looks okey. Any idea what other face swap I can use ?

People mention that to get superior upscale you have to use controlnet, so much confusion for such a simple thing!

I'm getting this error: InvalidArchiveError("Error with archive C:\\Users\\Administrator\\Downloads\ oop-unleashed-main\ oop-unleashed-main\\)

Hey, I need some help with this. I'm not sure how to make the eyes in the generated video look in the same direction as in the "Driving Video". When I generate the video, the eyes move like they do in the "Source Video". Any advice? Thank you in advance!

Hi, when I try to install it and run the bat I get this error on the cmd window: This script relies on Miniconda which cannot be silently installed under a path with spaces. All folders chained to roop don't have any spaces on their names example_example... How can I fix this? that prevents me from installing it? Thanks for the help, and great video.

why did this happen, why didn't I know this method 2 weeks ago, this is truly beyond reason and thank you for the knowledge🤝 I pray you stay healthy and your family too. "amin"🤲🏻

Completé la instalación pero e no puedo crear ninguna imagen porque me sale el error "CUDA out the memory". Tried to allocate 12.92 GiB. GUP". Me había animado a seguir este tutorial porque aquí decía que con 6GB de VRAM se podía, pero no. Será para otra vez o tendré que comprarme una GUP más potente 😁

Give me an unlocked model 😂 Holy shit, 30 mins to gen an image. That’s crazy. To get 3minutes I have to use control nets lol. FP8 or N for me. Even with a 3090 I still haven’t used FP16 but I should try it. I prefer having pruned models in unet. As you forget which include VAE or not.

It struggles with video when the subject is close and eating something. The image in the video is way off. Is there a way to correct that, with masking maybe?

very helpful video！I followed step by step, each step is explained clear and detailed. But I have no idea why it keeps showing errors (time out etc., very long msgs in red) when I do the syntax 04. Can u please help to point out possible issues, and the solutions?

conda envirnment is empty

how can i show svd , ive downloaded the model and it doesnt show up

It doesn't work. I added the command as "set COMMANDLINE_ARGS=--theme dark--xformers--no-half" and nothing happens, it still white.

I have checked and checked that everything is as you explained. However, My default aspect ratio does not automatically select 1920 X 1080. Any help?. Thanks for the tutorial.

How to use it on android tablet

website link ?

please tell me i wont have to use pinokio.

Sub'd Your videos are really helpful and cover all the little details other AI infulencers just speed/gloss over, Your pacing really helps too, so keep doing what you do! Now I wish we would get some flux control nets for Forge UI - I really need them! espeically open pose, reference, canny and seg!

One of the reasons some of the swapped faces look different is because the angle the heads are at are not the same as the image.

I have to warn people about the licensing. On the official release the licensing system for the images you generate are pretty lenient. Black Forest cannot come in and claim copyrights on you. The problem lies with all the other variants that have been made. You have to read the licensing if you plan on using the outputs for things like print on demands. Some people have provided very good models but have made the licensing so restrictive that you simply cannot use it for commercial usage. I am not joking. One of the best flux model just made me mad when i looked at licensing in Forge under the checkpoint tab, you can actually check the terms of licensing. I have see n many of the flux models asking to have attribution for the output or that the output cannot be used commercially. Be very careful about these models. SD3 is way better in terms of output licensing. Unless you generate for yourself than my wall of text means nothing but if you plan of using the imagery, be very careful.

Thanks for making the tutorial 🙌🏻

Good instructions, but I had an issue with loading safetensors. OmniGen\model.py:161: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requires_grad_(True), rather than torch.tensor(sourceTensor). self.register_buffer('scale_factor', torch.tensor(scale_factor)) C:\AI\MiniConda3\envs\env\lib\site-packages\diffusers\models\attention_processor.py:2358: UserWarning: 1Torch was not compiled with flash attention. (Triggered internally at ..\aten\src\ATen ative\transformers\cuda\sdp_utils.cpp:455.) hidden_states = F.scaled_dot_product_attention(