from transformers import AutoModels
model = AutoModels.from_pretrained("deepsynthbody/deepfake_ecg")
out = model(num_channels=5)