|
|
|
@ -145,7 +145,7 @@ and input(Y)
|
|
|
|
|
[0, 3, 6, 6, 8]]
|
|
|
|
|
ref_level: 0
|
|
|
|
|
then we get 1-level LoDTensor
|
|
|
|
|
Out.lod = [[0, 2, 5, 8]]
|
|
|
|
|
Out.lod = [[0, 1, 2, 5, 8]]
|
|
|
|
|
Out.data = [[a], [a], [b], [c], [d], [b], [c], [d]]
|
|
|
|
|
Out.dims = [8, 1]
|
|
|
|
|
|
|
|
|
@ -157,7 +157,7 @@ Given a common Tensor input(X)
|
|
|
|
|
and input(Y)
|
|
|
|
|
Y.lod = [[0, 2, 3, 6]]
|
|
|
|
|
ref_level: -1
|
|
|
|
|
then we a common Tensor
|
|
|
|
|
then we get a common Tensor
|
|
|
|
|
Out.data = [[a], [a], [b], [c], [c], [c]]
|
|
|
|
|
Out.dims = [6, 1]
|
|
|
|
|
|
|
|
|
|