LightningEnvironment¶

class pytorch_lightning.plugins.environments.LightningEnvironment[source]¶

Bases: pytorch_lightning.plugins.environments.cluster_environment.ClusterEnvironment

The default environment used by Lightning for a single node or free cluster (not managed).

The master process must be launched by the user and Lightning will spawn new worker processes for distributed training, either in a single node or across multiple nodes.

If the master address and port are not provided, the default environment will choose them automatically. It is recommended to use this default environment for single-node distributed training as it provides the most convenient way to launch the training script.

creates_children()[source]¶

Whether the environment creates the subprocesses or not.

Return type: bool

global_rank()[source]¶

The rank (index) of the currently running process across all nodes and devices.

Return type: int

local_rank()[source]¶

The rank (index) of the currently running process inside of the current node.

Return type: int

master_address()[source]¶

The master address through which all processes connect and communicate.

Return type: str

master_port()[source]¶

An open and configured port in the master node through which all processes communicate.

Return type: int

node_rank()[source]¶

The rank (index) of the node on which the current process runs.

Return type: int

teardown()[source]¶

Clean up any state set after execution finishes.

Return type: None

world_size()[source]¶

The number of processes across all devices and nodes.

Return type: int