[bug]: Unable to install a Flux CLIP model into Invoke

[kevcursor]

Is there an existing issue for this problem?

• I have searched the existing issues

Install method

Invoke's Launcher

Operating system

Linux

GPU vendor

Nvidia (CUDA)

GPU model

Nvidia 3080

GPU VRAM

10 GB

Version number

6.2.0

Browser

No response

System Information

No response

What happened

I have a CLIP model from Hugging face. I have tried a couple of them. When using "Scan Folder" or providing a local path it says "Model install error - Unable to determine model type." 6.1.0 said the same, but I haven't tried to install into any earlier version. CLIP Starter Models install fine, or at least they did when I tried a few versions ago, so I'm assuming this is a bug. Other models work as expected when I install via "Scan Folder."

What you expected to happen

I expected it to install like any other model but maybe there's an extra step I missed to specify which type of model it is.

How to reproduce the problem

Put CLIP model(s) into a directory, click "Scan Folder" and specify that directory. Click the "+" to the right of the listed model.

Additional context

Specifying a directory with image models works fine, as does specifying a directory with LoRa models etc. It fails for CLIP models and possibly for other types (not tested so I can't say). I have hunted through Google and through the GitHub issues etc., and nothing similar is showing.

Discord username

No response

[psychedelicious]

Please share a link to the model you are trying to install.

[kevcursor]

Thank you. The model I was trying to install was called "clip_l.safetensors" which can be found at https://huggingface.co/cyberdelia/FluxTextEnc_VAE/tree/main.