วีดีโอ

Hey man thanks for this video, For some reason the first generation is taking forever.. i left it for 2 hours and it was just at 5% and wasnt using my GPU at all was trying to use my CPU I dont know what could be the problem here

hi, I had my own Stable installation created before installing zluda, do you know if there is a way to start the one with zluda using some command without having to create a new version of Stable? Thanks :)

fatal: could not create work tree dir 'stable...': Invalid argument

Can roop unleashed & rope pearl be used on amd gpu? If yes then which amd gpu should i buy & what are the minimum requirements for amd gpu to run these ai applications? Waiting 4 ur reply..😊

it tell me that i have not enough ram but i have 16gb of ram. ?

i am getting this error - Exception Code: 0xC0000005 and it has a lot of lines from #0 to #116 (#114 0x00007ff7b7851494 __scrt_common_main_seh d:\agent\_work\2\s\src\vctools\crt\vcstartup\src\startup\exe_common.inl:288:0 #115 0x00007ffbe712257d (C:\WINDOWS\System32\KERNEL32.DLL+0x1257d) #116 0x00007ffbe832af28 (C:\WINDOWS\SYSTEM32 tdll.dll+0x5af28)) .

Anyone can help me remove zluda? I got rid of the AMD pro drivers and want to use the "normal" way. When i try setting up a new SD folder it always goes back and trys to use zluda.

are this tool support flux.1?

Thank you so much, finnally it worked!

I am getting this error - ImportError: cannot import name 'TypeIs' from 'typing_extensions' (D:\SDnext\automatic\venv\lib\site-packages\typing_extensions.py)

What does Replace ROCmLibs in Roblas Folder For Certain GPUs mean, I have a 6600XT and I'm not sure do I need to do anything? I don't even see a Roblas folder anywhere for that matter

"did not match any files known to git"

Dumb question, but do I just click run.bat when I want to start up SD Next again?

will this work with the 6800m? ive tried a few different techniques but it just wants to use the cpu

my laptop takes 1 hours for 20 seconds length, RAM 16GB and VRam 1GB, is that normal? or you have spesific settings for make it faster?

pls comfyui for zluda amd plsss

It is taking ages on 16GB Ram

How much time difference is there between 16GB and 32GB Ram?

Hi, when I try to run the program I get this error. RuntimeError: Found no NVIDIA driver on your system... How can i fix it? I have a AMD GPU and i wanted to try it on CPU

what the hell is "Fooocus" ? doesnt seems to have Loras options or anything interesting ?

I liked your narration, it was a fun video. Thank you very much.

hi , im getting ERROR: Exception in ASGI application , how will i fix it

I see in description about it work on AMD RX 6700m. Ummm... how?

Why is my speed at 1.9 it/s on a 7900 XTX.......??

running in zluda or directml ??? i dont see the message zluda detected on (your amd gpu)

THANK YOUUU

youre a life saver for amd user. i tried 3 times and finally succeses1!!!!!!!!!!!!!!!!!!!!!!!!!!

why I'm getting error with cuda stuff? sometimes it doesn't detect face images

does it mean it's better to use a faceset than just 1 image?.....

If you want to improve generation performance, and reduce to 5-7 steps while maintaining quality, I recommend using a Lora model ( lcm-lora-sdv1-5 ) You have to add "Add Network to Prompt": in UI settings you can add SD Lora to appear

Try this before anything else: Uninstall the python versions from your system and install only 3.10.6 (Add to Patch) This will also ensure that there are no errors during the process. Other versions with the Patch get in the way (at least it works for me)

Hey there I have the Problem that this Error accures RuntimeError: Torch is not able to use GPU; add --skip-torch-cuda-test to COMMANDLINE_ARGS variable to disable this check. In the comments another guy had the same question and you said: "make sure the hip ask you install is windows 10 & 11 5.7.1" What do you mean? Can you explain what I have to do? Great Video btw. Ok so I figgured that you ment the HIP SDK and mine was installed on WIndows 10 & 11 and it still didnt work, what did I do wrong :0

Very good. Now everything is running without errors, even Upscaling. Thanks you. Best Tutorial.

hello, i have the problem "ZLUDA device failed to pass basic operation test: index=None, device_name=AMD Radeon 'GPU' [ZLUDA]" and i did what you said to do: 1. venv\Scripts\activate 2. pip uninstall torch torchvision -y 3. webui-user.bat but i still have the problem,it says Using CPU-only torch and when i try to generate an image is just a black image and my graphic card driver keeps crashing. i have a rx 6700xt

1st comment! 🥰

can you explain what you did differently that it wont take 20-30 minutes for the first time ? because these steps look preety much the same as before

Upon executing the .bat script, I am redirected to Google Search, followed by a 127.0.0.1 error. I would appreciate any assistance.

Hey can you help with this? "The procedure entry point rocblas_dtrmm_outofplace could not be located in the dynamic link library F:\stable-diffusion-webui-amdgpu\.zluda\cublas64_11.dll." It's a windows pop up when I tried to run webui-user.bat with --use-zluda on. I clicked ok and the web ui loads but it uses CPU instead. In the command line there is also this "Failed to load ZLUDA: [WinError127] The specified procedure could not be found"

Whats the difference using faceset vs 1 picture?

3:00 where can I get those "model check points"? And what are they?

Hey bro, I am using faceset of like 10 photos but in some parts of videos, the face like blurs which make it somewhat clear that is generated by Ai, any solutions to that, btw I am using max front photos in the faceset and 4 photos from a little here and there angle

Hey bro , I can't use DML due to low VRAM but I can use CPU in roop unleashed, is that a problem on the quality of output or it's just that it takes more time?

Hi My memory usage is very low only using 2 gb and generating 2 frames per second. I have more vram and ram . How to utilize them. I tried using codeformer only. Also in command prompt the execution thread is showing as 1 only. Mine have 16 threads I tried increasing it in the settings but not worked

I tried this method 4 times. For some reason I followed the exact steps but I was getting errors. I tried another model instead of juggernaut xl. It is currently downloading the 6.2gb file in the command line when I ran the run.bat file 🥹

Once everything is done, its giving me error -- Hmmm… can't reach this page127.0.0.1 refused to connect. What can be the issue?

Greetings, could you also make an example with an Nvidia card and if it were possible for a low-resource PC?

C: oop-unleashed-4.0.0>git pull fatal: not a git repository (or any of the parent directories): .git

Why this eror appear D:\a\_work\1\s\onnxruntime\core\providers\cuda\cuda_call.cc:121 onnxruntime::CudaCall D:\a\_work\1\s\onnxruntime\core\providers\cuda\cuda_call.cc:114 onnxruntime::CudaCall CUDA failure 35: CUDA driver version is insufficient for CUDA runtime version ; GPU=-2133276112 ; hostname=DESKTOP-KS33I53 ; file=D:\a\_work\1\s\onnxruntime\core\providers\cuda\cuda_execution_provider.cc ; line=245 ; expr=cudaSetDevice(info_.device_id);

Donate me at Kofi: ko-fi.com/tongtong Paypal: tongtong83@gmail.com Related Video: th-cam.com/video/0p-mCnhTDNA/w-d-xo.html My PC's Spec Intel Core i5 10400F AMD Radeon RX 6600 XT 16GB RAM Video: A Young Kid Playing The Ukelele by cottonbro studio www.pexels.com/video/a-young-kid-playing-the-ukelele-3682076/ Chapters 0:00 - Intro 0:05 - Update Roop-Unleashed 0:27 - Link to Roop-Unleashed If You Haven't Install It 0:44 - Setup the DML in Provider 1:04 - Preparing Photos For Training 1:46 - Faceset Training 2:36 - Face Swap Using Trained Faceset Roop Unleashed has just been updated to version 4.1.1. If you already installed it previously, you can update it by going to the roop-unleashed folder, typing CMD in the address bar, and pressing Enter. From there, type git pull, press Enter again, and you are good to go. To run Roop Unleashed, double-click on the `run.bat` file. However, if you haven't installed it, you can watch this video on how to install it. The link will be at the top right of the video or in the description and comments below. Please note that this tutorial is specifically made for AMD GPUs only; other GPUs like Nvidia will not work. Right now, I'm going to show you how to train a faceset for future use in Roop Unleashed. Before we begin, go to the Settings, scroll down to Provider, and choose dml from the drop-down list. Click Apply Settings and Restart Server. This ensures that it will use DirectML throughout the operation. For this tutorial, I will be using Bruce Lee's face as a demo. This is strictly for educational purposes and should not be used for any illegal operations. As you can see, I have prepared 21 photos of Bruce Lee to train his faceset. There is no rule on how many faces are needed, but I personally suggest that you prepare at least 10 photos. Make sure there are no other foreign faces in the photos, as this will affect the final outcome. In this photo, you can see two blurred faces in the background; while these won't be detected, it is still advisable to crop the photo just to be safe. Go to the “Face Management” at the top of the web UI. Scroll down until you see "Input Files." Select all the photos prepared earlier and drag them into it. After a while, all the faces that have been detected will be shown as thumbnails at the bottom. Go through each face and make sure there are no foreign faces mixed in. You can always select the photo and click "Remove selected" to remove the unwanted face. Once you are satisfied, you can click "Create/Update Faceset file." A faceset file will be created on the left side of the screen. You can download the file and rename it to something easy to remember. I saved mine as Bruce Lee. Drag the Bruce Lee faceset to the Source Files and the video you want to face swap to the Target Files. I chose Codeformer from the drop-down list in "Select post-processing." You may choose anything you prefer. The newest version has a new option called "Subsample upscale to." Higher numbers will give a clearer face. I chose 256 pixels for this tutorial. Finally, scroll all the way down and click "Start" to begin. As you can see, the conversion utilizes around 20% of CPU, 10 GB of RAM, and 60% of GPU. I'm showing this only for your reference as I have a few viewers spreading hate and scolding me because they claimed that it fully utilizes their CPU. Just saying, cursing me won't help with anything. Everyone has different PC specs and all of them work differently. It took me around 7 minutes and 50 seconds to convert a 21-second video. The original video size is 14.2 MB and after the conversion, the file size has become 36.7 MB. You can always find your converted file in the roop-unleashed, output folder. Please note that if your video has a lot of weird angles, the faceswapped result may not be perfect. I hope you enjoy the video and please consider donating to me so I can make more useful videos for you in the future.

can run Forge webui with Zluda?