StreamDiffusionTD 0.1.9 🚀🖼️ [ TOX ] [ works on Mac ]
Only a few changes but exciting ones :
Mac (Apple Silicon) supported ! (Will need capable apple silicon ! slower than nvidia cards but working )
Model Preset Config system added - allows for all Models to be set with a preset... (handy for recalling tensorrt setups)
Big shoutout / thank you to @barretto for the help getting this working cross platform.
For mac, if you change resolution or restart stream, it might have trouble finding the buffer and you will see a printout in the terminal. A workaround for now is changing stream out name or disable/enable the operator.
Comments (36)
Has a solution been found for cuda? I tried it yesterday and can't get any further. Can someone help me? Regards, Micha
which cuda version are you having trouble with?
But the Cuda toolkit isnt available for Mac, I'm confused
did anyone find a way to get it to work on Mac? I have it running on my PC but it's in for repairs... Any help is greatly appreciated
impossible to make it work...
I'm struggling to get this working on my M1 Mac. The install instructions seem to be just for Windows. Any update on tutorials for this?
Hi, I've been experimenting with this this Tox for StreamDiffusion and really like it; except that the images that I'm getting are far from great (txt to img), no matter how well defined and detailed is the prompt that I am using. SD Turbo model works best that others but is still far from good images. Any advice?
y tambien me da error el instalar el tensorRT
casi lo tengo pero me salta este error cuando le doy a start:Last login: Thu Aug 8 13:54:06 on ttys000 /Users/macintosh/Documents/IA/StreamDiffusion/Start_StreamDiffusion.sh ; exit; macintosh@MacBook-Pro-de-Macintosh ~ % /Users/macintosh/Documents/IA/StreamDiffusion/Start_StreamDiffusion.sh ; exit; Traceback (most recent call last): File "/Users/macintosh/Documents/IA/StreamDiffusion/streamdiffusionTD/main_sdtd.py", line 17, in from torchvision.transforms.functional import to_tensor ModuleNotFoundError: No module named 'torchvision' Saving session... ...copying shared history... ...saving history...truncating history files... ...completed. [Proceso completado]
/Desktop/TDAI/StreamDiffusion/streamdiffusionTD/main_sdtd.py", line 17, in from torchvision.transforms.functional import to_tensor ModuleNotFoundError: No module named 'torchvision' did someone resolve this ?
Is there maybe a tutorial on how to get it working on Mac ? thanks
Does this only run on Apple Silicon macs? not intel?
i really wanna know how to make it work on my mac
How to install CUDA Toolkit on the mac? Need help
If someone who has successfully got it up and running on Mac would be willing to make an install video, I'm sure there are plenty of people who would be immensely grateful.
I installed it successfully. I will DM DotSimulate to see if I can make a tutorial for it
Hi, it seems I have everything well but I get this error message on the StreamDiffusionTD tox: error: (parameter:seed) td.tdAttributeError: 'td.baseCOMP' object has no attribute 'curBlock'
I've tried with a random value for the seed so I've no error anymore. But now if I start the stream i've "script errors inside network"
i'm stuck with python version 3.10 not installed. i tried with venv and miniconda
First of all thank you so much for the amazing work! I just downloaded Stream Diffusion 0.1.9 (for Mac) but having hard time installing it. I’m running MacBook Pro, 2023 (M2 Max, 96GB, Ventura 13.5) Touchdesigner ver. 2023.11760 (16bit) Python 3.12.3 Downloaded tox and opend it. Download/Update SD: (finished and folders were created, though I didn’t see a command prompt open) Install SD (venv + all req): clicked on it but nothing happend. Could you tell me what I did wrong? Thank you so much in advance!
I would realllly love to get sdxl-turbo running. So far with the only change being changing the hugging face model ID to sdxl-turbo from sd-turbo FPS goes from 6 to .002. Been consistent on multiple computers. Hope you have insight, thx DS!
Hey all, I am having trouble installing on mac using the directions in the install tab, Can you direct me to how can anyone help?
did you install it ?
hello there. I have a problem when opening the .tox fie inside TD in the parameter box the settings tab is empty
That is likely due to incompatible TD version. It needs to be a build after the official 2023 production build (a few experimental builds dont load it). or would be my best guess
Hellooo Anyone has this running on Apple silicon yet? My setup: python 3.10.9 macOS 14.1 - M3 - 39GB TouchDesigner 2023.11760 NDI SDK v6 (not sure if this is even necessary? I'm running into some "problems": 1. First of all, in any case, I always get a warning when opening StreamDiffusionTD 0.1.9 in Touchdesigner (No idea if this is related specifically to my setup? Anyone else?): " TouchDesigner argument error An unknown argument: -B was given to TouchDesigner as a command argument. If a path to a .toe file is being given, ensure it has quotes around it if it has spaces in the path. Exiting. " 2. If I install everything using the plugin, I get the missing CUDA when executing "Install SD (venv + all req) - I'm just ignoring that and it will install anyway. 3. When starting the stream, I only get one static image, it won't change/update on changing inputs, etc.. I tried different installation method (installing stream diffusion manually, Pytorch using "https://pytorch.org/get-started/locally/" from: https://pytorch.org/get-started/locally/ etc..) once got to the point where it started generating multiple images but now after trying something else, its back to one static output :D So I messed it up again. When it works I get around 1.6 FPS - is that around about what I can expect from my setup? If so there is no particular reason to further explore this approach in my opinion - I'll rather use a separate machine with nvidia then ;) In any way, thanks for putting in all the work, I really appreciate it and enjoy using the plugin otherwise !
Having the same issue with unknown argument
Hi, is it possible to create two independent sd in the same project?
Hi, I am so happy to try it out thanks for the huge work and sharing! I also have problems with the running file. Is there any installation guide also for the Mac available?
Hi, i have installed everything and when i try to test stream it, shows error "No module named 'torchvision'"
also when i installed the environment it showed error:"ModuleNotFoundError: No module named 'pip' [end of output] note: This error originates from a subprocess, and is likely not a problem with pip. ERROR: Failed building wheel for nvidia-pyindex Failed to build nvidia-pyindex ERROR: Could not build wheels for nvidia-pyindex, which is required to install pyproject.toml-based projects Installing dependencies with pip from the activated virtual environment... Requirement already satisfied: pip in d:\streamdiffusion\streamdiffusion\venv\lib\site-packages (24.0) Looking in indexes: https://download.pytorch.org/whl/cu118 ERROR: Could not find a version that satisfies the requirement torch==2.1.0 (from versions: 2.2.0+cu118, 2.2.1+cu118, 2.2.2+cu118, 2.3.0+cu118) ERROR: No matching distribution found for torch==2.1.0
Hey! Already having so much fun with your work, would love a link to join your discord! Just have few issues with ComfyTD, SD working amazing though 👍🏼 Thanks
check the about page and there is a link that will open if you hit About Me pulse, there should be a link available to discord server.
When I try to install Stream Diffusion, I always get the same error at "pulse install SD." It says "Cuda detection failed" even though I've installed Cuda 11.8 several times by now. Can you please help me out with this problem? Cheers.
You should be good if you restart PC or TouchDesigner !
Been waiting for this! So excited.