Lyell Hintz patreon

StreamDiffusionTD 0.2.0🚀🖼️

🕑 Added 2024-08-18 16:02:00 +0000 UTC

Comments

Airelle Dubois

Hello, I don't know if this is the good place to ask for help, let me know if there is actually an other. While attempting to start stream, a window signal "ModuleNotFoundError: No module named 'cv2'". What can I do to repair this? Thank you

MIkhail Loginev

i hve an isue. I cant instal Failed to activate virtual environment. Please check the path and ensure the venv exists.

MIkhail Loginev

same isue

Luciana Maia

Hey there, do you know how can i solve this problem? Current directory: C:\Users\lumaia86\Documents\Touchdesigner\STREAM DIFFUSION TEST 2.2\StreamDiffusion Changed directory to: C:\Users\lumaia86\Documents\Touchdesigner\STREAM DIFFUSION TEST 2.2\StreamDiffusion Creating Python venv at: "C:/Users/lumaia86/Documents/Touchdesigner/STREAM DIFFUSION TEST 2.2/StreamDiffusion\venv" C:\python-3.10.11-embed-amd64\python.exe: No module named venv Attempting to activate virtual environment... El sistema no puede encontrar la ruta especificada. Failed to activate virtual environment. Please check the path and ensure the venv exists. Path to venv: "C:/Users/lumaia86/Documents/Touchdesigner/STREAM DIFFUSION TEST 2.2/StreamDiffusion\venv" VIRTUAL_ENV: "" Presione una tecla para continuar . . .

MoritzW

Downgrading huggingface-hub to version 0.24.6 fixed it for me.

Eduardo Ortega

same issue here

Andreas

same issue

Vj Igor Cardoso

mine gives this error when I click on the pulse of the star stream: Traceback (most recent call last): File "M:\Ai_1\StreamDiffusion\streamdiffusionTD\wrapper_td.py", line 683, in _load_model pipe = pipeline_class.from_pretrained( File "M:\Ai_1\StreamDiffusion\venv\lib\site-packages\diffusers\pipelines\pipeline_utils.py", line 1090, in from_pretrained cached_folder = cls.download( File "M:\Ai_1\StreamDiffusion\venv\lib\site-packages\diffusers\pipelines\pipeline_utils.py", line 1856, in download cached_folder = snapshot_download( File "M:\Ai_1\StreamDiffusion\venv\lib\site-packages\huggingface_hub\utils\_validators.py", line 114, in _inner_fn return fn(*args, **kwargs) File "M:\Ai_1\StreamDiffusion\venv\lib\site-packages\huggingface_hub\_snapshot_download.py", line 216, in snapshot_download raise LocalEntryNotFoundError( huggingface_hub.errors.LocalEntryNotFoundError: Cannot find an appropriate cached snapshot folder for the specified revision on the local disk and outgoing traffic has been disabled. To enable repo look-ups and downloads online, pass 'local_files_only=False' as input. During handling of the above exception, another exception occurred: Traceback (most recent call last): File "M:\Ai_1\StreamDiffusion\streamdiffusionTD\wrapper_td.py", line 687, in _load_model except huggingface_hub.utils._errors.LocalEntryNotFoundError: AttributeError: module 'huggingface_hub.utils' has no attribute '_errors' Failed to load model due to: module 'huggingface_hub.utils' has no attribute '_errors'

Frédéric Bussière

is it possible to see just the result of the control-net, for example the openpose or the hed? just to make sure this part works as expected. thanks, love your work

fidlar

This is brill! However I keep getting 'status: server connected, not streaming' do you know what this could be? Thanks!

Jeroen Bocken

Not sure who needs this, but Python 3.12 gave me some errors. Just use the recommended python 3.10.9. Once I changed, the installation went perfect.

Mark Jones

Thanks for this and the guide. I've been having some wierd behaviour with it, I am running it on a 4090 system with a Threadripper 3970X and am getting approx 15fps with default settings, I have a 2nd PC with a 4070 and a 7700x cpu and am getting 16fps, any ideaswhy I'm getting better performance with the cheaper build?

Wesley Messina

Any updates on when the next update is coming. And your outlook on what’s to come?

McPixel

It's no problem to do that. You have to play around with noise and adding that by using a comp node to get the noise onto your webcam video input. :-)

Drew Molleur

any updates on this? our (un)stable diffusion keeps crashing haha

Andrew Hlynsky

not working for me :(

Iven

Just change the Resolution in Setting and after that change the streamout text in setting 2 and it will run. This is a workaround.

Andrew Hlynsky

Still struggling to change the resolution. I'm just getting started, so I don't have any projects to rebuild. If you feel like sending me the file i would be incredibly thankful!

Eugene O'Donald

Hay, thanks for the solution. I keep running into this as soon as I activate acceleration though. Is there a workaround for that to? EDIT: I was using a pretty odd resolution 1920x546 I think that was the reason it did not work, staying with "normal" resolutions seems to work!

Stoyan Stoyanov

Has anyone tried successfully to use another model then the SD Turbo. I cannot understand how to get it to work, since it errors when downloading from hugging face id, and the local ones are not visible in the drop down

McPixel

Hey, thank you very much for that great plugin! Do you know when the resolution fix get's online? Thank you! Best Korbi

BreakItDown

No Mac support?

BreakItDown

Mac please

Veronika Shiliak

Hello! I've followed tutorial to download stable diffusion. I don't have any errors appearing, but the output is just black and SD doesn't seem to work. Do you know what could be the problem? Did anyone faced something similar? Thank you!

Enviral Design

hey thanks for your work on this, is there a path forwards to load civit ai models into streamdiffusionTD?

梓王

hello, i just bought it. i would like to know that is it can be used commercially?

Iven

Hey, awesome work. On my Laoptop everything is fine. But i try to run it on my desktop and i always get this after installing when i try to start streaming in my cmd.exe (...path To RepostreamdiffusionTD\main_sdtd.py:485: SyntaxWarning: invalid escape sequence '\C' print(f"\Controlnet Stream '{input_mem_name} not found. Try changing the Stream Out Name (to SD) parameter in Stream Settings 2 of TD operator.") Sorry i dont get it. Need some help please. Thank You!

Charles Coragem

Christopher Lensing

Just to complete it here (as you also answered on discord but it‘s perhaps good to know): it seems to be linked to the models version but I did test LoRas trained on 2.1 that worked and some that didn‘t. I‘ll keep on testing and meanwhile use the working ones, which I am fine with.

dotsimulate

Likely due to model mismatch. What sd model are you using? You will need to make sure the LoRA is trained for that model pipe (1.5,2.1,sdxl, etc)

Christopher Lensing

I get an error concerning LoRas: ERROR loading LoRA Models: Error(s) in loading state_dict for UNet2DConditionModel: size mismatch for down_blocks.0.attentions.0.proj_in.lora_A.WillemVanAelst.weight: copying a param with shape torch.Size([128, 320, 1, 1]) from checkpoint, the shape in current model is torch.Size([128, 320]). size mismatch for down_blocks.0.attentions.0.proj_in.lora_B.WillemVanAelst.weight: copying a param with shape torch.Size([320, 128, ... (error message shortened by 63 lines). Any idea why?

dotsimulate

Yes that is known bug with this version. it works if you change the streamoutname after changing the resolution but I also have a fix that I am planning on posting today. Trying to get an updater op finalized so replacing it doesn't require project rebuild at all.

Riverdhan

can you do one with live cam being generated as ai, i want to use it for live cam visual

Iven

is there a way to use a different resolution than 512x512 like in previous version?

João Rodrigues

Having trouble to connect to Discord and having problems with Creating Python venv

dotsimulate

can you send me a DM (discord ideally just for screenshots,etc or patreon) with more info on this? The operator copies code into the folder each time the stream starts so if you edited and saved the file itself that would actually have no difference. If you changed the code in the operator in order to fix this, this is definitely something for me to take note of.

dotsimulate

Hit the Install 2 pulse and it will install that library. You are just missing a single additional library that was added in this version.

dotsimulate

Check your firewall or network settings and make sure python.exe or TouchDesigner is not blocked or restricted.

Matthieu Tourne

It was working on mac a couple of versions ago .. to make this one do anything I had to change the hardcoded 'cuda' references in main_sdtd.py:370 ``` input_tensor = torch.empty((3, height, width), dtype=torch.float32, device=device) control_input_tensor = torch.empty((3, height, width), dtype=torch.float32, device=device) pinned_input_buffer = torch.empty((height, width, 3), dtype=torch.uint8, pin_memory=False) pinned_control_buffer = torch.empty((height, width, 3), dtype=torch.uint8, pin_memory=False) ```

jianyu mo

A matching Triton is not available, some optimizations will not be enabled. Error caught was: No module named 'triton' Traceback (most recent call last): File "C:\Users\JAM\StreamDiffusion\streamdiffusionTD\main_sdtd.py", line 37, in from wrapper_td import StreamDiffusionWrapper File "C:\Users\JAM\StreamDiffusion\streamdiffusionTD\wrapper_td.py", line 24, in from pipeline_td import StreamDiffusion File "C:\Users\JAM\StreamDiffusion\streamdiffusionTD\pipeline_td.py", line 19, in from v2v_utils import get_nn_latent #V2V File "C:\Users\JAM\StreamDiffusion\streamdiffusionTD\v2v_utils.py", line 5, in from einops import rearrange ModuleNotFoundError: No module named 'einops' why~