Forge UI - non-AIO Flux models don't pick up options for text_encoders #1103
Closed
routesmith
started this conversation in
General
Replies: 3 comments 1 reply
|
The fix was a simple one: Add the full path of your text_encoder directory to the Stable Diffusion WebUI Forge as additional arguments: Example: This will allow WebUI Forge to get the CLIP / T5 options in the dropdown for non-AIO Flux models. HTH |
0 replies
|
You shouldn't have to do that if you use |
1 reply
|
In A1111 this is defined as "--clip-models-path". Is there any documentation for the dirs and paths in Forge, to get the difference?? |
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
It seems the non-AIO Flux models don't pick up the CLIP / T5 text encoder options from the "Data/Models/text_encoder" directory. Is this the correct place to store those files for the UI to pick them up as options, or should they go somewhere else? Also, I had to create the text_encoder directory under "Data/Models" and noticed that there was a "Data/Packages/stable-diffusion-webui-forge/models/text_encoder" that didn't seem to carry over for whatever reason.
I searched pretty extensively and only found reference to possibly breaking VAE / text encoder directory config options you can pass on startup, with no resolution. Any pointers would be welcome!
Thanks!
All reactions