Argmax

Description

Returns the index corresponding to the maximum value on the specified axis.

Input

One input:

data: Input tensor, the data type supports float16, float, and int32, and the data format supports ND.

Output

One output:

y: tensor of type int64, indicating the index position of the maximum value. The dimension is one less than that of the input x.

Property

axis: (required) int32, specifying the direction for calculating the maximum index. The value for axis is [-len(x.shape), len(x.shape)-1].

keep_dim: (optional) either 1 (default) or 0.

select_last_index: indicates whether to select the first or last position when the maximum value occurs at multiple positions. The value can be 0 or 1. The default value is 0, and only 0 is supported currently.

ONNX Opset Support

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