i dont suggest it anymore
i dont suggest it anymore
Finstall directly
Fi have shown in several videos recently
FYouTubeSECourses
Welcome to the Pillars of AI Application Libraries: How to Install Tutorial! 
Are you ready to embark on a journey of installing AI libraries and applications with ease? 
In this video, we'll guide you through the process of installing Python, Git, Visual Studio C++ Compile tools, and FFmpeg on your Windows 10 machine. 
We'll also show you ho...
Are you ready to embark on a journey of installing AI libraries and applications with ease? In this video, we'll guide you through the process of installing Python, Git, Visual Studio C++ Compile tools, and FFmpeg on your Windows 10 machine. We'll also show you ho...Fand this one
FYouTubeSECourses
Stable Diffusion Gets A Major Boost With RTX Acceleration. One of the most common ways to use Stable Diffusion, the popular Generative AI tool that allows users to produce images from simple text descriptions, is through the Stable Diffusion Web UI by Automatic1111. In today’s Game Ready Driver, NVIDIA added TensorRT acceleration for Stable Diff...
Fwhat is your GPU?
SThat is an A5000 on runpod. Running via cli though, rather than automatic1111, so I wonder if there is some settings that I'm missing. i.e. perhaps in automatic 1111, doubling the batch size will automatically half the total number of training steps?
GThanks for reply, i fixed just now. The error was because i didnt install insightface.
FA5000 cant run multiple batch size
Fbatch size 1 already uses 17 gb
Fyes it requires c++
Shmmm interesting, i haven't received an out of memory error on that with a batch size of 2
MHey guys, since I have to make an impression on the influencers based on a real one, I should slightly change their face with stable diffusion. In your opinion, how could it be done?
Fthen something must be wrong
Sokay gotcha. Yeah, I guess that should use double the amount of vram, exceeding what the a5000 can do
F40 gb card can do 3 batch size
F48 gb can do 4
Fi dont remember testing 2 on 24 
Sokay gotcha, thanks for that info
MI was thinking about controlNet, for you am I on the right way?
S@Dr. Furkan Gözükara when you use a batch size of 4 on a 48gb card, did you find you needed to adjust your learning rate?
Fye may work
Fi use same
Fbut i didnt test actually
Fhowever it works
Fi trained several times
SGuys any of you knows if I upload my data to chat gpt is possible that they take it to re train their model? or someone else get the data trough a question that someone can ask?
Syes I know. @Dr. Furkan Gözükara please let us know if there is another channel to ask this type of questions without interfere with stable difussion questions
STheir terms claim that any data you send into their APIs directly will not be used for training. Using ChatGPT (the web UI), however, the data will be used for training in the future https://community.openai.com/t/where-can-i-read-the-terms-and-conditions-for-the-api-service/365738
EThese are amazing :Fire:
VSo i just tried an a6000 with 48GB memory, using your 48GB_TextEncoder.json. Tried to set batch size higher than 1, but I get an out of memory error on 4, 3, or 2, so I had to keep it at 1. Doesn't seem any faster than 4090 with batch size 1.
RMerry Christmas to everyone
RI am new to join here
RI am trying to create my own SDXL lora using diffusers
RBut not working well
RDoes anyone can help me?
TYou can try creating dreambooth
RYeah, I am trying to
RWhat I am trying to do is using diffusers to create a dreambooth lora model
Fye with that you have to enable xformers
Fwelcome
Fwhat help you need
MThis is only 8 bit and yet it's highly accurate, thank you!
Fyes llava is very powerful
MWhat GPU do you suggest on runpod for 16bit? 3090 is not enough
RI wanna create SDXL lora model on my own images, but not working well.
MI want Sd WebUi API text2Image for multiple user so we can track every user process percentage runtime, Is there any extension or other repo availble for ready to used
Fi think go with a6000
F48 gb