Fine-tuning original ControlNet Models

[basit-MT]

Is it possible to Fine-tune original ControlNets instead of fine-tuning your own from scratch using SD as base models?

For example, can I perhaps, further Fine-tune "lllyasviel/control_v11f1p_sd15_depth" for my own Custom data? Is this supported with Diffusers?

[tolgacangoz]

Is setting this argument what you are looking for:

diffusers/examples/controlnet/train_controlnet.py

Lines 261 to 267 in 214990e

	parser.add_argument(
	"--controlnet_model_name_or_path",
	type=str,
	default=None,
	help="Path to pretrained controlnet model or model identifier from huggingface.co/models."
	" If not specified controlnet weights are initialized from unet.",
	)

[basit-MT]

Thanks @tolgacangoz, completely overlooked it.