mindspore_gl.nn.GNNCell

class mindspore_gl.nn.GNNCell[source]

GNN Cell class.

Construct function will be translated by default.

Supported Platforms:

Ascend GPU

static disable_display()[source]

Disable display code comparison.

Examples

>>> from mindspore_gl.nn import GNNCell
>>> GNNCell.disable_display()
static enable_display(screen_width=200)[source]

Enable display code comparison.

Parameters

screen_width (int, optional) – Determines the screen width on which the code is displayed. Default: 200.

Examples

>>> from mindspore_gl.nn import GNNCell
>>> GNNCell.enable_display(screen_width=350)
classmethod sparse_compute(csr=False, backward=False)[source]

Whether to use sparse operator to accelerate calculation.

Parameters
  • csr (bool, optional) – Is it a csr data structure. Default: False.

  • backward (bool, optional) – Whether to use custom back propagation. Default: False.

Raises

ValueError – If csr is False and backward is True.

Examples

>>> from mindspore_gl.nn import GNNCell
>>> GNNCell.sparse_compute(csr=True, backward=False)
classmethod specify_path(path)[source]

Enable specify the construct file path.

Parameters

path (str) – The path for save the construct file.

Examples

>>> from mindspore_gl.nn import GNNCell
>>> GNNCell.specify_path('path/to/save')