How to combine Inpaint and ContentRef templates?

Hi, thanks for this great project.

I am trying to use these two templates together:

• DiffSynth-Studio/Template-KleinBase4B-Inpaint
• DiffSynth-Studio/Template-KleinBase4B-ContentRef

My goal is:

• Use Inpaint to modify only the masked area.
• Use ContentRef to guide the generated content/style.

Both templates work individually, but I am not sure how to combine them correctly.

I tried loading both templates:

template = TemplatePipeline.from_pretrained(
    torch_dtype=torch.bfloat16,
    device="cuda",
    model_configs=[
        ModelConfig(model_id="DiffSynth-Studio/Template-KleinBase4B-Inpaint"),
        ModelConfig(model_id="DiffSynth-Studio/Template-KleinBase4B-ContentRef"),
    ],
)

and passing both inputs:

template_inputs=[
    {
        "model_id": 0,
        "image": image,
        "mask": mask,
        "force_inpaint": True,
    },
    {
        "model_id": 1,
        "image": reference_image,
    },
]

But I got:

paste error or describe the unexpected result here

1. Is combining Inpaint + ContentRef supported?
2. Is this the correct way to pass multiple templates?
3. Are there any additional settings required for ContentRef + Inpaint?

Thanks!

Start with the TemplatePipeline.from_pretrained entry point and the ModelConfig and template_inputs usage shown in the issue. Check whether Inpaint and ContentRef are supported together and what the completed input should be; done means a reproducible supported example or a clear limitation and error explanation.