- class lightning.pytorch.plugins.io.TorchCheckpointIO¶
This is an experimental feature.
- load_checkpoint(path, map_location=<function TorchCheckpointIO.<lambda>>)¶
Loads checkpoint using
torch.load(), with additional handling for
fsspecremote loading of files.
- Return type:
Returns: The loaded checkpoint.
FileNotFoundError – If
pathis not found by the
Remove checkpoint file from the filesystem.
- save_checkpoint(checkpoint, path, storage_options=None)¶
Save model/training states as a checkpoint file through state-dump and file-write.