ReduceMin

Description

Computes the minimum value of the input Tensor elements along the specified axis. If keepdims equals 1, the resulting Tensor has the same number of dimensions as the input Tensor. If keepdims equals 0, the resulting Tensor removes the reduced dimensions.

Input

data: input tensor of type uint8, int32, float16, or float32. The data format is ND.

Output

reduced data：output tensor, whose data type is the same as that of the input data.

Property

axes: The data type is an int list. Meaning: calculation axis. Value range: [-r, r-1], where r is the number of dimensions of the input data x.

keepdims: int, specifying whether to retain the reduced dimensions. Defaults to 1.

Constraints

When axes is empty, full-dimensional reduction is performed on all axes by default. In this case, it is recommended to modify the axes of the ONNX operator to all axes; or before exporting the ONNX graph from PyTorch, use amin in the PyTorch model to perform reduction on all axes (for example, x.amin(dim=[0, 1, 2])).

ONNX Opset Support

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