How can I used a fine-tuned a VAD segmentation 3.0 checkpoint?

[gabitza-tech]

Hello,

When I try to use a fine-tuned VAD segmentation 3.0 checkpoint like this:

model = Model.from_pretrained(
  "model.ckpt", 
  use_auth_token="HUGGINGFACE_ACCESS_TOKEN_GOES_HERE")
vad_pipeline = VoiceActivityDetection(segmentation=model)
HYPER_PARAMETERS = {
                            "min_duration_on": min_on,
                            "min_duration_off": min_off,
                            }
vad_pipeline.instantiate(HYPER_PARAMETERS)

out_vad = vad_pipeline({"audio": filepath})

I get:

RuntimeError: A pipeline must be instantiated with `pipeline.instantiate(parameters)` before it can be applied.

Could you please help?