Expand

Description

Broadcasts the input tensor following the given shape and the broadcast rule.

Input

• x: input tensor of type float16, float, bool, int64, int32, int8, or uint8. The data format is ND.
• shape: input tensor of type int64.

Output

y: output tensor of the identical data type as x. The shape of y must be derived from the input shape of x. The data format can be ND.

Constraints

You need to modify the model to change the input shape from placeholder to const. You can use ONNX Simplifier to simplify the model.

ONNX Opset Support

Opset v8/v9/v10/v11/v12/v13/v14/v15/v16/v17/v18