EmbeddingBag

Description

Computes the inverse output of the embedding function.

Input

Two mandatory inputs and two optional inputs:

weight: (required) tensor of type float.

indices: (required) tensor of type int32.

offset: (optional) tensor of type int32.

per_sample_weights: (optional) tensor of type float.

Output

One output:

y: tensor of type float.

Property

mode: string type.

Scale_grad_by_freq: bool type.

sparse: bool type.

include_last_offset: bool.

ONNX Opset Support

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