tf.experimental.dtensor.create_mesh
Stay organized with collections
Save and categorize content based on your preferences.
Creates a single-client mesh.
tf.experimental.dtensor.create_mesh(
mesh_dims: Optional[List[Tuple[str, int]]] = None,
mesh_name: str = '',
devices: Optional[List[str]] = None,
device_type: Optional[str] = None
) -> tf.experimental.dtensor.Mesh
If both mesh_dims and devices are specified, they must match each otehr.
As a special case, when all arguments are missing, this creates a 1D CPU mesh
with an empty name, assigning all available devices to that dimension.
Args |
|---|
mesh_dims
|
A list of (dim_name, dim_size) tuples. Defaults to a single
batch-parallel dimension called 'x' using all devices. As a special case,
a single-element mesh_dims whose dim_size is -1 also uses all devices.
|
mesh_name
|
Name of the created mesh. Defaults to ''.
|
devices
|
String representations of devices to use. This is the device part
of tf.DeviceSpec, e.g. 'CPU:0'. Defaults to all available logical devices.
|
device_type
|
If devices is missing, the type of devices to use. Defaults
to 'CPU'.
|
Returns |
|---|
|
A single-client mesh created from specified or default arguments.
|
Except as otherwise noted, the content of this page is licensed under the Creative Commons Attribution 4.0 License, and code samples are licensed under the Apache 2.0 License. For details, see the Google Developers Site Policies. Java is a registered trademark of Oracle and/or its affiliates. Some content is licensed under the numpy license.
Last updated 2023-03-17 UTC.
[null,null,["Last updated 2023-03-17 UTC."],[],[],null,["# tf.experimental.dtensor.create_mesh\n\n\u003cbr /\u003e\n\n|-------------------------------------------------------------------------------------------------------------------------------|\n| [View source on GitHub](https://github.com/tensorflow/tensorflow/blob/v2.9.3/tensorflow/dtensor/python/mesh_util.py#L47-L110) |\n\nCreates a single-client mesh. \n\n tf.experimental.dtensor.create_mesh(\n mesh_dims: Optional[List[Tuple[str, int]]] = None,\n mesh_name: str = '',\n devices: Optional[List[str]] = None,\n device_type: Optional[str] = None\n ) -\u003e ../../../tf/experimental/dtensor/Mesh\n\nIf both `mesh_dims` and `devices` are specified, they must match each otehr.\nAs a special case, when all arguments are missing, this creates a 1D CPU mesh\nwith an empty name, assigning all available devices to that dimension.\n\n\u003cbr /\u003e\n\n\u003cbr /\u003e\n\n\u003cbr /\u003e\n\n| Args ---- ||\n|---------------|--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|\n| `mesh_dims` | A list of (dim_name, dim_size) tuples. Defaults to a single batch-parallel dimension called 'x' using all devices. As a special case, a single-element mesh_dims whose dim_size is -1 also uses all devices. |\n| `mesh_name` | Name of the created mesh. Defaults to ''. |\n| `devices` | String representations of devices to use. This is the device part of tf.DeviceSpec, e.g. 'CPU:0'. Defaults to all available logical devices. |\n| `device_type` | If `devices` is missing, the type of devices to use. Defaults to 'CPU'. |\n\n\u003cbr /\u003e\n\n\u003cbr /\u003e\n\n\u003cbr /\u003e\n\n\u003cbr /\u003e\n\n| Returns ------- ||\n|---|---|\n| A single-client mesh created from specified or default arguments. ||\n\n\u003cbr /\u003e"]]
View source on GitHub