Shrink

Description

Performs non-linear transformation on the input tensor and applies scales and bias to the input using the bias based on the relationship between the input value data and the threshold lambda.

Input

data: input tensor of type float16 or float.

Output

y: output tensor of the identical data type and shape as input data.

Attribute

• bias: float. Defaults to 0.0.
• lambda: float. Defaults to 0.5.

ONNX Opset Support

Opset v9/v10/v11/v12/v13/v14/v15/v16/v17/v18