|TensorFlow 1 version||View source on GitHub|
Deletes old checkpoints.
Compat aliases for migration
See Migration guide for more details.
tf.train.CheckpointManager( checkpoint, directory, max_to_keep, keep_checkpoint_every_n_hours=None, checkpoint_name='ckpt' )
Used in the notebooks
|Used in the guide||Used in the tutorials|
import tensorflow as tf checkpoint = tf.train.Checkpoint(optimizer=optimizer, model=model) manager = tf.train.CheckpointManager( checkpoint, directory="/tmp/model", max_to_keep=5) status = checkpoint.restore(manager.latest_checkpoint) while True: # train manager.save()
CheckpointManager preserves its own state across instantiations (see the
__init__ documentation for details). Only one should be active in a
particular directory at a time.
tf.train.Checkpointinstance to save and manage checkpoints for.
directory: The path to a directory in which to write checkpoints. A special file named "checkpoint" is also written to this directory (in a human-readable text format) which contains the state of the
max_to_keep: An integer, the number of checkpoints to keep. Unless preserved by
keep_checkpoint_every_n_hours, checkpoints will be deleted from the active set, oldest first, until only
max_to_keepcheckpoints remain. If
None, no checkpoints are deleted and everything stays in the active set. Note that
max_to_keep=Nonewill keep all checkpoint paths in memory and in the checkpoint state protocol buffer on disk.
keep_checkpoint_every_n_hours: Upon removal from the active set, a checkpoint will be preserved if it has been at least
keep_checkpoint_every_n_hourssince the last preserved checkpoint. The default setting of
Nonedoes not preserve any checkpoints in this way.
checkpoint_name: Custom name for the checkpoint file.
checkpoints: A list of managed checkpoints.
Note that checkpoints saved due to
keep_checkpoint_every_n_hourswill not show up in this list (to avoid ever-growing filename lists).
latest_checkpoint: The prefix of the most recent checkpoint in
directoryis the constructor argument to
Suitable for passing to
tf.train.Checkpoint.restoreto resume training.
max_to_keepis not a positive integer.
save( checkpoint_number=None )
Creates a new checkpoint and manages it.
checkpoint_number: An optional integer, or an integer-dtype
Tensor, used to number the checkpoint. If
None(default), checkpoints are numbered using
checkpoint.save_counter. Even if
save_counteris still incremented. A user-provided
checkpoint_numberis not incremented even if it is a
The path to the new checkpoint. It is also recorded in the