Video suggestion: How to make positive and negative prompts consistently, and how to choose the best configuration for generating images (sampling method, Schedule type, sampling steps, CFG scale), and everything else needed to generate quality images.
great video...just a note...if you already have Miniconda you should download and install the latest version..or you might have problems that cause installation to not be able to set up the environment..after updating Miniconda the process went fairly smoothly...but be patient with the installation...no connection errors occur when downloading models...just re-enter the syntax and try again..(information packets failing to download..internet speed ect. responsible not the program..once again i entered the syntax and successfully installed Omnigen..thx again for the tutorial..
@@FoxtonAI ..Thank you for all your effort and research..it has been very informative and concise..one question I have, if you are able to answer for me is how do you access the "remove blur" selector of do i have to prompt it a certain way..the same with images created on my p.c...it does not seem to recognize it and produces a rendering error..any help would be greatly appreciated..Thanks once again and keep up the great work!
I don't quite have 6 GB vram, only have 4 GB vram but do have around 50 GB Ram. My GPU is using the latest Nvidia driver, though. Will this be doable on my system or am I out of luck since I'm basically 2 GB shy of 6 GB of vram?
Hello. I have a graphics card with 12GB of memory and I have installed the FP32 version. And even worked a little bit in OmniGen. Today, when creating some images, there were messages about a lack of memory. I want to reinstall the 8-bit version. What do I need to remove from the installed one and what steps from this video should I follow to install another version?
Just checked the syntax below again...and it works fine. Did you make sure to add TWO underscores after the question mark? Add this to the end of your local URL to enable dark mode: /?__theme=dark Adding the dark theme to the run.bat file is not as easy as you would think, there are technical complexities with the operations. Hope that helps.
I cant not keep consistent animated animals on flus nor A1111, only human characters can be consistent, do you know if I could keep consistent animated animals to create a story?
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(
My Nvidia GPU isn't used in generating the image. It is always 0%. I have an Intel IRIS Xe GPU (comes with the CPU), but that is being used instead. That's not optimal, is it?
Perfect step by step video.
Thank you! Glad you liked the video.
Through tutorial, very concise, thank you kindly.
Thank you for your comment. Glad you liked the video.
Video suggestion: How to make positive and negative prompts consistently, and how to choose the best configuration for generating images (sampling method, Schedule type, sampling steps, CFG scale), and everything else needed to generate quality images.
Thanks for that suggestion. I may take a look at doing exactly that.
great video...just a note...if you already have Miniconda you should download and install the latest version..or you might have problems that cause installation to not be able to set up the environment..after updating Miniconda the process went fairly smoothly...but be patient with the installation...no connection errors occur when downloading models...just re-enter the syntax and try again..(information packets failing to download..internet speed ect. responsible not the program..once again i entered the syntax and successfully installed Omnigen..thx again for the tutorial..
Thank you for taking the time to make all those useful points, they'll be very helpful for others. Glad you liked the video.
@@FoxtonAI ..Thank you for all your effort and research..it has been very informative and concise..one question I have, if you are able to answer for me is how do you access the "remove blur" selector of do i have to prompt it a certain way..the same with images created on my p.c...it does not seem to recognize it and produces a rendering error..any help would be greatly appreciated..Thanks once again and keep up the great work!
very nice tutorial. thanks!
Thank you! Glad you liked the video.
All god now! Thank you
Glad you got it sorted.
Instructions very clear and have loaded unfortunately after processing an image after about 10 mins stops showing error message, any ideas
Glad the video was helpful. That's a bit strange. Could be a number of things with your setup. Check for clues with the error message.
I have GeForce GTX 1650 . Which is not cuda capable. So can I install this without cuda.?
You could do, but it will probably use your CPU instead. You could give it a go.
help, it doesn't recognize the [SYNTAX 02]: in the cmd
Hope you got this issue sorted.
I don't quite have 6 GB vram, only have 4 GB vram but do have around 50 GB Ram. My GPU is using the latest Nvidia driver, though. Will this be doable on my system or am I out of luck since I'm basically 2 GB shy of 6 GB of vram?
You could test it, but for this 8-bit version, I think you’re going to need at least 6GB.
Hello. I have a graphics card with 12GB of memory and I have installed the FP32 version. And even worked a little bit in OmniGen. Today, when creating some images, there were messages about a lack of memory. I want to reinstall the 8-bit version. What do I need to remove from the installed one and what steps from this video should I follow to install another version?
Although, I only have 16GB of RAM - maybe it's not enough...
Hello. I would install and use the 8-bit version. How did you install the FP32 version, so I know what to suggest.
Well, I tried that but it opens google search in browser. I thought I could add this command inside the run bat file or something
Just checked the syntax below again...and it works fine.
Did you make sure to add TWO underscores after the question mark?
Add this to the end of your local URL to enable dark mode:
/?__theme=dark
Adding the dark theme to the run.bat file is not as easy as you would think, there are technical complexities with the operations.
Hope that helps.
nice tutorial. how to add the theme dark to startup?
Add this to the end of your local URL:
/?__theme=dark
whats the status on a rocm version?
Unfortunately, not heard anything for AMD compatibility.
I cant not keep consistent animated animals on flus nor A1111, only human characters can be consistent, do you know if I could keep consistent animated animals to create a story?
Which video are you referring to? This video is for installing OmniGen.
@ on omnigen
my run.bat not working even the directory is changed to the folder i installed..kindly help please
That's strange. If you paste the content of your run.bat file here, I can take a quick look.
@@FoxtonAI i can start manually by right click the folder and start a terminal, then insert command conda activate omnigen, so on so on
@@mohdhazwan7706 if you have spaces in your folder path you may need to put quotes around the path.
can i use that on rtx3050 4gb vram?
You could always test it, but I think you will find you need a minimum of 6GB VRAM.
Help me with omnin control please
Depending on what you're trying to achieve, take a look at some specific videos available.
please tell me i wont have to use pinokio.
You won't have to use Pinokio.
Have you got a video for people who know nothing? You lost me at GPU
Unfortunately not. I know it can all be a bit daunting for beginners, but watch it a couple of times, I literally show every step on screen.
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(
My Nvidia GPU isn't used in generating the image. It is always 0%. I have an Intel IRIS Xe GPU (comes with the CPU), but that is being used instead. That's not optimal, is it?
Strange. Did you get it sorted?
You will need to use your NVIDIA GPU.