Split the api to support comfyui like workflow

[chinshou]

For instance, some effects require combining multiple different control networks to achieve them, which is easy to implement in ComfyUI. However, this is not possible with the current Stable Diffusion C API. Is it feasible to split the API into modules like module loader, VAE decoder/encoder, latent image generation, k-sampler, control network application, etc.?

[pedroCabrera]

I'm also looking into this, I have successfully managed to split model loading and inferencing with the scheduler, rgn, and all the parameters reset on fast inference, but now looking into split it more as you said and also wondering about control net merging.