openvino.runtime.opset15.embedding_bag_packed#

openvino.runtime.opset15.embedding_bag_packed(emb_table: Node | int | float | ndarray, indices: Node | int | float | ndarray, per_sample_weights: Node | int | float | ndarray | None = None, reduction: Literal['sum', 'mean'] = 'sum', name: str | None = None) → Node#

Return a node which performs sums or means of “bags” of embeddings, without the intermediate embeddings.

Parameters:

  • emb_table – Tensor containing the embedding lookup table.

  • indices – 2D Tensor of shape [batch, indices_per_bag] with indices.

  • per_sample_weights – Tensor of weights to be multiplied with embedding table with same shape as indices.

  • reduction – Operator to perform reduction of elements in bag.

  • name – Optional name for output node.

Returns:

The new node performing EmbeddingBagPacked operation.