TorchCheckpointIO

class pytorch_lightning.plugins.io.TorchCheckpointIO[source]

CheckpointIO that utilizes torch.save() and torch.load() to save and load checkpoints respectively, common for most use cases.

load_checkpoint(path, map_location=<function TorchCheckpointIO.<lambda>>)[source]

Loads checkpoint using torch.load(), with additional handling for fsspec remote loading of files.

Parameters

path (Union[str, Path]) – Path to checkpoint
map_location (Optional[Callable]) – a function, torch.device, string or a dict specifying how to remap storage
locations. –

Returns: The loaded checkpoint.

Raises: FileNotFoundError – If path is not found by the fsspec filesystem
Return type: Dict[str, Any]

remove_checkpoint(path)[source]

Remove checkpoint file from the filesystem.

save_checkpoint(checkpoint, path, storage_options=None)[source]

Save model/training states as a checkpoint file through state-dump and file-write.

Parameters

checkpoint (Dict[str, Any]) – dict containing model and trainer state
path (Union[str, Path]) – write-target path
storage_options (Optional[Any]) – not used in TorchCheckpointIO.save_checkpoint

Raises

TypeError – If storage_options arg is passed in

Return type

None