tf.cumsum

tf.cumsum(
    x,
    axis=0,
    exclusive=False,
    reverse=False,
    name=None
)

Defined in tensorflow/python/ops/math_ops.py.

See the guide: Math > Scan

Compute the cumulative sum of the tensor x along axis.

By default, this op performs an inclusive cumsum, which means that the first element of the input is identical to the first element of the output:

tf.cumsum([a, b, c])  # [a, a + b, a + b + c]

By setting the exclusive kwarg to True, an exclusive cumsum is performed instead:

tf.cumsum([a, b, c], exclusive=True)  # [0, a, a + b]

By setting the reverse kwarg to True, the cumsum is performed in the opposite direction:

tf.cumsum([a, b, c], reverse=True)  # [a + b + c, b + c, c]

This is more efficient than using separate tf.reverse ops.

The reverse and exclusive kwargs can also be combined:

tf.cumsum([a, b, c], exclusive=True, reverse=True)  # [b + c, c, 0]

Args:

• x: A Tensor. Must be one of the following types: float32, float64, int64, int32, uint8, uint16, int16, int8, complex64, complex128, qint8, quint8, qint32, half.
• axis: A Tensor of type int32 (default: 0). Must be in the range [-rank(x), rank(x)).
• exclusive: If True, perform exclusive cumsum.
• reverse: A bool (default: False).
• name: A name for the operation (optional).

Returns: