search

invoke-ai / InvokeAI

Invoke is a leading creative engine for Stable Diffusion models, empowering professionals, artists, and enthusiasts to generate and create visual media using the latest AI-driven technologies. The solution offers an industry leading WebUI, and serves as the foundation for multiple commercial products.
https://invoke-ai.github.io/InvokeAI/
Apache License 2.0
23.34k stars 2.4k forks source link

[bug]: Can we get a one time Install thing ? "caught was: No module named 'triton'" #4166

Closed LubuLubu2 closed 5 months ago

LubuLubu2 commented 1 year ago

Is there an existing issue for this?

OS

Windows

GPU

cuda

VRAM

No response

What version did you experience this issue on?

3.0.1

What happened?

Error caught was: No module named 'triton'

Can we get some Extract and Run Portable version ? InvokeAI and WebUI are killing me. How many times i have to Download ~3GB Torch ? how many times i have to wait almost an hour for Toch to Install while the whole PC is forzen and all i can do is to watch an Ugly cmd window ? Why are there so many downloads ? i specifically said no model download and uncheked all of them but i downloaded +4GB data anyways. and after all of that "Error caught was: No module named 'triton'" happened. doesn't matter if i ignore it or run another command, i get a bunch of errors and another "Error caught was: No module named 'triton'" loop.

Download ~3GB Torch... Wait almost an hour for Torch... Download a bunch of unwanted Models, one ~2GB, Another Torch whatever 1.7GB, and another Torch dot whatever around 1.5GB.... Everything is set to go... Error caught was: No module named 'triton'.... Run re-Install.... re-Installing Torch while everything is frozen... Same Error....

Screenshots

No response

Additional context

No response

Contact Details

No response

Millu commented 1 year ago

Could you share the terminal output once you hit [1] to start using Invoke?

mohamedaleya commented 1 year ago

Just had this error too I'm still installing the models though. Will update this comment once everything is setup. I do agree with you that it would be great to have a Portable version as an alternative.

WindowsTerminal_3CSsxcNniu

Millu commented 1 year ago

@mohamedaleya @LubuLubu2 you should be able to ignore the Triton error message on windows - try going ahead and using the WebUI and see if you run into any issues

LubuLubu2 commented 1 year ago

@Millu Millu Sorry man i deleted Invoke and don't have the files anymore. Option 1 to 10 were the same, none of them did anything, same loop.

FilipTrichkov commented 1 year ago

I have the same issue as well when I launch invoke.bat, I can still generate images in the WebUI but it is signifcantly slower. I thought this was an issue because I did not update from 2.3.5 to 3.0.1 properly but I still can generate images so I am not sure. I tried installing triton in my new Invoke 3 directory with the 'pip install triton' command, but I was left with this message "ERROR: Could not find a version that satisfies the requirement triton (from versions: none) ERROR: No matching distribution found for triton"

Screenshot 2023-08-07 160441
mohamedaleya commented 1 year ago

@Millu Millu

Sorry man i deleted Invoke and don't have the files anymore. Option 1 to 10 were the same, none of them did anything, same loop.

Same here. Ended up deleting entirely cause I kept getting an error saying no module named "invokeai" after downloading about 32gb of files, whenever I press any option in the menu.

bigsk1 commented 1 year ago

Same error fresh install, webui works but... errors bug me. Windows 11, zip file install. Python 3.10.11, cuda toolkit, 12.2, latest drivers, C++ libraies, I also run 1111, Comfy Ui, no errors.

Please enter 1-10, Q: [1] 1 Starting the InvokeAI browser-based UI.. A matching Triton is not available, some optimizations will not be enabled. Error caught was: No module named 'triton' [2023-08-18 23:19:09,046]::[InvokeAI]::INFO --> Patchmatch initialized X:\InvokeAI\InvokeAI-Installer\invoke.venv\lib\site-packages\torchvision\transforms\functional_tensor.py:5: UserWarning: The torchvision.transforms.functional_tensor module is deprecated in 0.15 and will be removed in 0.17. Please don't rely on it. You probably just need to use APIs in torchvision.transforms.functional or in torchvision.transforms.v2.functional. warnings.warn( [2023-08-18 23:19:11,647]::[uvicorn.error]::INFO --> Started server process [23512] [2023-08-18 23:19:11,650]::[uvicorn.error]::INFO --> Waiting for application startup. [2023-08-18 23:19:11,652]::[InvokeAI]::INFO --> InvokeAI version 3.0.2post1 [2023-08-18 23:19:11,663]::[InvokeAI]::INFO --> Root directory = X:\InvokeAI\InvokeAI-Installer\invoke [2023-08-18 23:19:11,671]::[InvokeAI]::INFO --> GPU device = cuda NVIDIA GeForce RTX 4090 [2023-08-18 23:19:11,721]::[InvokeAI]::INFO --> Scanning X:\InvokeAI\InvokeAI-Installer\invoke\models for new models [2023-08-18 23:19:11,975]::[InvokeAI]::INFO --> Scanned 5 files and directories, imported 0 models [2023-08-18 23:19:12,017]::[InvokeAI]::INFO --> Model manager service initialized [2023-08-18 23:19:12,045]::[uvicorn.error]::INFO --> Application startup complete. [2023-08-18 23:19:12,046]::[uvicorn.error]::INFO --> Uvicorn running on http://127.0.0.1:9090 (Press CTRL+C to quit)

Harvester62 commented 8 months ago

Both messages are not a problem under Windows. Simply ignore them. Therefore, if these were the only reported problems, they are not a bug per se, so we might consider to close this bug report.