Concat

tensorflow::ops::Concat

Concatenates tensors along one dimension.

Summary

Arguments:

• scope: A Scope object
• values: List of N Tensors to concatenate. Their ranks and types must match, and their sizes must match in all dimensions except concat_dim .
• axis: 0-D. The dimension along which to concatenate. Must be in the range [-rank(values), rank(values)).

Returns:

• Output: A Tensor with the concatenation of values stacked along the concat_dim dimension. This tensor's shape matches that of values except in concat_dim where it has the sum of the sizes.

Concat block

Source link :https://github.com/EXPNUNI/enuSpaceTensorflow/blob/master/enuSpaceTensorflow/tf_array_ops.cpp

Argument:

• Scope scope : A Scope object (A scope is generated automatically each page. A scope is not connected.)
• Input values: List of N Tensors to concatenate. Their ranks and types must match, and their sizes must match in all dimensions except concat_dim .
• StringPiece axis: 0-D. The dimension along which to concatenate. Must be in the range [-rank(values), rank(values)).

Return:

• Output output : Output object of Concat class object.

Result:

• std::vector(Tensor) result_output : The output tensor.

Using Method

※ values에 shape와 type이 모두 같은 tensor리스트가 들어오고, axis에는 리스트의 tensor를 통합할 차원을 선택한다. 위 화면의 예시 처럼 {2,2}의 shape를 가진 3개의 tensor를 리스트에 모으고, axis에 첫번째 차원을 합치기 위해 0(index번호로 0은 첫번째라는 의미)이라는 값을 입력한다. 이렇게 하면 결과 값은 {6,2}의 shape를 가진 tensor가 나온다.