Output model type on the [ Load Diffusion Model ] node

[JorgeR81]

Feature Idea

Add a model type output to the Load Diffusion Model node. The model type could then be connectected to the Dual Clip Loader node.

Existing Solutions

Set the type option manually, in the Dual Clip Loader.

Other

I'm sure this would be useful in other scenarios. So a type output for the Load Checkpoint node could also be useful.

And if you don't want to make the current nodes more complicated, you could just create "advanced" versions.

[ltdrdata]

Strictly speaking, the diffusion model and CLIP are completely separate. It's possible to not use the model at all, load the CLIP with DualClipLoader for text encoding, then unload the CLIP and separately utilize only the generated conditioning

[JorgeR81]

So, you mean that this is not a practical solution ?

Because we could have a scenario where the CLIP encoder is loaded before the UNET model, and then the "model type" would not be available ?

[ltdrdata]

They are just completely independent. The model loader cannot represent the functionality of the clip loader.