[bug]: i cant install sd 3.5 large turbo

[tvortsa]

Is there an existing issue for this problem?

• I have searched the existing issues

Operating system

Windows

GPU vendor

Nvidia (CUDA)

GPU model

rtx3060

GPU VRAM

6Gb

Version number

5.4.1

Browser

chrome

Python dependencies

No response

What happened

i cant install sd 3.5 large turbo
i get error: Cannot determine base type

What you expected to happen

I expexted for model will be install

How to reproduce the problem

No response

Additional context

No response

Discord username

No response

[Anothergazz]

Im having the same problem with SD3.5 large
I'm using InvokeAi V5.4.2 within Stability Matrix. When I use 'Add Model' -> 'scan folder' Invoke finds the file but fails to import it.

[REDDKOMET]

Same issue but I have no issues with 3.5 large and medium, just turbo can't be imported. Works fine in Comfy.

[tampadesignr]

using without matrix just invoke itself i had to have my huggingface token setup first

[Anothergazz]

using without matrix just invoke itself i had to have my huggingface token setup first

Thanks for the suggestion, I had already set it up, and I have just checked on Hugging face and it looks OK

[freelancer2000]

Do they support Safetensors format for SD 3.5? What they install from the model manager that they have setup are the Diffusers version.

[tampadesignr]

Do they support Safetensors format for SD 3.5? What they install from the model manager that they have setup are the Diffusers version.

it looks like a clone of the hugginface repo with whatever is needed to be a diffuser.. lots of files.. not completely sure just looking at both folders
https://huggingface.co/stabilityai/stable-diffusion-3.5-large/tree/main/transformer
think thats the diffuser part