@@VJSCHOOL I tried building a model with different res and more steps, but it wouldn't finish... Super impressed with the speed, it's still super jittery and "stabley", but that's just animdif in general I geuss, excited to try out control net!!
Thanks for this great solution! Amazing performance on 4070 12GB! SD Turbo and 1 sample step. I will test making other engine now. 3 sample steps were slow on 4070.
THIS video talks about the portable version and NOT the manual installation. If you try following this installation for the manual installation, you are going to face errors related to models and engines not being found, for example "cannot access local variable 'pipe' where it is not associated with a value", because the paths for the portable version and the manual version are slightly different. That's why I recommend using the portable version
Hello, first of all congratulations on the tremendous work I have been using the DSD patch for a long time without any problems, but now I have a constant error that I cannot solve As soon as I open the project it tells me that I have a problem with the line 4 of the TouchDiffusionExt operator that does not find the Torch module within the project...
Nevertheless I got finished engine preparation and selected the file which webui includes, it doesnt work.. I tried to press load engine and refresh engine several times, but it dosent work.. how to resolve this :(
I don't understand how it's posssible to have multiple frames per second with 20 steps. Why don't we use TensorRT on normal UI's like Automatic or Comfy ? Sorry in advance for having such a vague question but i find this quite amazing haha.
Real-time can be achieved with Turbo models and Acceleration Lora, on low steps like 1-4 with TensorRT. TouchDesigner allow to copy data from and to GPU, that decrease latency.
excuse me, i follow every step you said, but when i click the webui button, it was going to fail, would you tell me why? here are the flows: Downloading StreamDiffusion... Cloning into 'StreamDiffusion'... remote: Enumerating objects: 2735, done. remote: Counting objects: 100% (1115/1115), done. remote: Compressing objects: 100% (294/294), done. error: RPC failed; curl 92 HTTP/2 stream 5 was not closed cleanly: CANCEL (err 8) error: 7789 bytes of body are still expected fetch-pack: unexpected disconnect while reading sideband packet fatal: early EOF fatal: fetch-pack: invalid index-pack output Failed to download StreamDiffusion
OSError: [WinError 126] The specified module could not be found. Error loading "C:\Users\PC ABU-ABU\Desktop\TouchDiffusion-main\TouchDiffusion-main\.venv\Lib\site-packages\torch\lib\fbgemm.dll" or one of its dependencies. No launch file found please help
Very interesting, but when I clicked load engine in touchdesigner, I got an error like "cannot access local variable 'pipe' where it is not associated with a value". What is the problem?
@@VJSCHOOL Tried the portable as well as the manual version. Same error. It may be because Runway shut down their SD1.5 repositories on huggingface? Posted the full log on GitHub Issue. Thanks!
hi ,i have followed your tutorial and every step went well .But in TouchDesigner,the log shows :cannot access local variable 'pipe' where it is not associated with a value.I don't know how did it happened
the webui.bat did not create any folders (venv, streamdiffusion, models, engines) but tells me they already exist after relaunching, what am I doing wrong?
the .venv folder is hidden so make sure you have Show Hidden Folders turned on in Windows Explorer. It didn't create the models/engines directories for me either, I just copied and pasted them manually from the StreamDiffusion folder once it was generated and that did the trick.
Git is installed. Creating .venv directory... The filename, directory name, or volume label syntax is incorrect. Failed to create virtual environment. What is the problem?
1) everything installed in venv, so it not affecting main Python env. 2) if you want to try with Python 3.10, then provide path to it in webui & use TouchDesigner version before 2023
amazing fast speeds would you know why using my null of the render tops of my td projects as input isnt working? im getting flashing white and black with some cat images but noise top works normally, denoising value dosent change this would we have to repeat the webui model proceses with higher steps seems like it could be that or a different problem if so how should i do that thanks!
@@VJSCHOOL yeah I got it to run stable diffusion but it was quite slow, and then I couldn’t figure out how to put it in touch designer, I’ll just bite the bullet return this computer and get one with an nvidia videographics card
@@VJSCHOOL I fixed this error but again, when I click on the "make engine" button, It does not work and a simple "error" appears without any info about it.
Thanks for making this free, much love
amazing job explaining the entire process...
instnaly fast and so nice to be in 512, congrats on this one, cant wait for more features!
Custom resolutions already available, but keep in mind, that it will affect FPS.
@@VJSCHOOL I tried building a model with different res and more steps, but it wouldn't finish...
Super impressed with the speed, it's still super jittery and "stabley", but that's just animdif in general I geuss, excited to try out control net!!
this is great! lean and clean, thank you!
Thanks for sharing your knowledge , i appreciate it 🎉
Thanks for this great solution! Amazing performance on 4070 12GB! SD Turbo and 1 sample step. I will test making other engine now. 3 sample steps were slow on 4070.
Amazing
THIS video talks about the portable version and NOT the manual installation. If you try following this installation for the manual installation, you are going to face errors related to models and engines not being found, for example "cannot access local variable 'pipe' where it is not associated with a value", because the paths for the portable version and the manual version are slightly different. That's why I recommend using the portable version
Hi, I've downloaded the portable version when i use the webui.bat, should I still set the python location to my local file? thanks
Thanks bro
Hello, first of all congratulations on the tremendous work
I have been using the DSD patch for a long time without any problems, but now I have a constant error that I cannot solve
As soon as I open the project it tells me that I have a problem with the line 4 of the TouchDiffusionExt operator that does not find the Torch module within the project...
Nevertheless I got finished engine preparation and selected the file which webui includes, it doesnt work.. I tried to press load engine and refresh engine several times, but it dosent work.. how to resolve this :(
thank you so much! great tutorial. is it possible to use other kind of
models?
does this work with MacOS Silicone?
I was wondering the same thing...
I was able to run streamdiffusion by it self, works great
Okay. So basically this utilizes a tensorrt engine to render in Touchdesinger. This needs an Nvidia card so no for a mac.
hi, i followed every step and there are no erros, but in touchdesigner when i hit pulse on "load engine" nothing happens, can anybody help me?
Create issue on GitHub or Discord and share full log.
Did you find the solution? Same issue
I don't understand how it's posssible to have multiple frames per second with 20 steps. Why don't we use TensorRT on normal UI's like Automatic or Comfy ? Sorry in advance for having such a vague question but i find this quite amazing haha.
Real-time can be achieved with Turbo models and Acceleration Lora, on low steps like 1-4 with TensorRT. TouchDesigner allow to copy data from and to GPU, that decrease latency.
I have tried it.. but say try experimenting with my video camera will it not work?
Thank you ! I made it work but it just creates the image only when i press load engine button and its a stable image. How can i make it live =?
excuse me, i follow every step you said, but when i click the webui button, it was going to fail, would you tell me why?
here are the flows:
Downloading StreamDiffusion...
Cloning into 'StreamDiffusion'...
remote: Enumerating objects: 2735, done.
remote: Counting objects: 100% (1115/1115), done.
remote: Compressing objects: 100% (294/294), done.
error: RPC failed; curl 92 HTTP/2 stream 5 was not closed cleanly: CANCEL (err 8)
error: 7789 bytes of body are still expected
fetch-pack: unexpected disconnect while reading sideband packet
fatal: early EOF
fatal: fetch-pack: invalid index-pack output
Failed to download StreamDiffusion
where i prerating models?
OSError: [WinError 126] The specified module could not be found. Error loading "C:\Users\PC ABU-ABU\Desktop\TouchDiffusion-main\TouchDiffusion-main\.venv\Lib\site-packages\torch\lib\fbgemm.dll" or one of its dependencies.
No launch file found
please help
Hey! Did u find the answer? I have the same problem
Very interesting, but when I clicked load engine in touchdesigner, I got an error like "cannot access local variable 'pipe' where it is not associated with a value". What is the problem?
That’s mean either model or engine can’t be found
Have you solved it?
@@yanyu813 download portable version on Github and try with it.
@@VJSCHOOL Tried the portable as well as the manual version. Same error. It may be because Runway shut down their SD1.5 repositories on huggingface?
Posted the full log on GitHub Issue.
Thanks!
I got this error when i click on Make Engine "UnboundLocalError: cannot access local variable 'pipe' where it is not associated with a value"
same here. Were you able to solve it?
@@joaopedrofireman8180 Yes i have installed the portable version without any error
hi ,i have followed your tutorial and every step went well .But in TouchDesigner,the log shows :cannot access local variable 'pipe' where it is not associated with a value.I don't know how did it happened
Create issue on GitHub or Discord with full log
thats amazing but i got an ERROR: ERROR: Failed to build installable wheels for some pyproject.toml based projects (tensorrt) please help!
"No module named 'tensorrt'"
As I understand it, this miracle cannot be run under Mac?
Only PC & Windows
amazing... this is realy faster,
can we put lora and vae to the folder?
what lora and vae for the best creation
and why icant change the seed ?
Tiny VAE already baked in for best performance. Loras not supported yet.
Seed can be changed in TouchDesigner with seed parameter
the webui.bat did not create any folders (venv, streamdiffusion, models, engines) but tells me they already exist after relaunching, what am I doing wrong?
the .venv folder is hidden so make sure you have Show Hidden Folders turned on in Windows Explorer. It didn't create the models/engines directories for me either, I just copied and pasted them manually from the StreamDiffusion folder once it was generated and that did the trick.
норм
How to increase CFG to 7-10 instead of 1?
The quality of generations is very low at 1.
You should try to increase sampling steps (batch size)
its posible with 8gvram?
Git is installed.
Creating .venv directory...
The filename, directory name, or volume label syntax is incorrect.
Failed to create virtual environment.
What is the problem?
chech where the python with "where python" command and copy paste in to webui.bat like he said
you need to install python first; if you don't have it - it won't be able to create virtual environment
Hellllooooo, I just arrive untill 4:05 and this happen : NameError: name 'StreamDiffusionWrapper' is not defined
Any idea how to fix this ?
There is portable version on GitHub, better to try with it
HOW MUCH FPS YOU GET MY FRIEND
does it have to be python 311? i have 310 and i don't wanna break all my other installations to try this...
1) everything installed in venv, so it not affecting main Python env.
2) if you want to try with Python 3.10, then provide path to it in webui & use TouchDesigner version before 2023
What's the difference between this and streamdiffusion?
Explained in first 3 min of video
It’s near x2 faster
Cool! Can it work with controlnet?
No support for now
amazing fast speeds would you know why using my null of the render tops of my td projects as input isnt working? im getting flashing white and black with some cat images but noise top works normally, denoising value dosent change this would we have to repeat the webui model proceses with higher steps seems like it could be that or a different problem if so how should i do that thanks!
seems the same problem with a default moveifileintop
I figured out I had to reinitalize pulse touchddiffusion folder and possibly clear log too I did both then reload engine
I am getting an error: name 'StreamDiffusionWrapper' is not defined. Any ideas?
same over here :( did you solve it?
Try portable version
any progress with portable version? I can not download it
I really need a computer to do this. is there a way I could offload that GPU onto Google Collab?
TouchDiffusion component required TouchDesigner, so it can’t run in Google Collab. You can try with original StreamDiffusion repo.
@@VJSCHOOL - Sigh guess i just need better hardware.
@@VJSCHOOLsweet I’ll try that
@@VJSCHOOL I got a new computer with a 3060 Ti =D
Can you do this with AMD Ryzen Somehow?
For now there is no way
@@VJSCHOOL yeah I got it to run stable diffusion but it was quite slow, and then I couldn’t figure out how to put it in touch designer, I’ll just bite the bullet return this computer and get one with an nvidia videographics card
@@SwoleKitchen yep 😄
Hi! Does it also work with Mac?
Only NVIDIA GPU
how about mac~ pls❤
Only NVIDIA GPU
anybody test 4080 laptop fps?
4070 Laptop - 24 FPS with SD Turbo 512x512
Faster than stream diffusion?
I get this error "Error installing packages: torch==2.1.0, Failed to install TensorRT" during the installation process. Any ideas how to fix this?
Try with portable version
@@VJSCHOOL I fixed this error but again, when I click on the "make engine" button, It does not work and a simple "error" appears without any info about it.