VariableSaver¶

class tfsnippet.VariableSaver(variables, save_dir, max_versions=2, filename='variables.dat', latest_file='latest', save_meta=True, name=None, scope=None)¶

Bases: tfsnippet.utils.reuse.VarScopeObject

Version controlled saving and restoring TensorFlow variables.

Attributes Summary

`name`	Get the name of this object.
`variable_scope`	Get the variable scope of this object.

Methods Summary

`get_latest_file`()	Get the latest available checkpoint file.
`restore`([ignore_non_exist])	Restore the checkpoint from file if it exists.
`save`([global_step])	Save the checkpoint to file.

Attributes Documentation

name¶: Get the name of this object.

variable_scope¶: Get the variable scope of this object.

Methods Documentation

get_latest_file()¶: Get the latest available checkpoint file.

restore(ignore_non_exist=False)¶

Restore the checkpoint from file if it exists.

Parameters:	ignore_non_exist (bool) – Whether or not to ignore error if the checkpoint file does not exist? (default `False`)
Raises:	`IOError` – If the checkpoint files do not exist, and ignore_non_exist is not `True`.

save(global_step=None)¶

Save the checkpoint to file.

Parameters:	global_step (int or tf.Tensor) – The global step counter.