Caffe2 - Python API
A deep learning, cross platform ML framework
Public Member Functions | List of all members
caffe2.python.checkpoint.CheckpointManager Class Reference
Inheritance diagram for caffe2.python.checkpoint.CheckpointManager:

Public Member Functions

def __init__ (self, db_prefix, node_name, db_type, metadata_handler=None)
 
def init (self, nodes=None, retrieve_from_epoch=None, path_prefix=None, path_type=None)
 
def blob_list (self)
 
def load (self, epoch, path_prefix=None, path_type=None)
 
def load_blobs_from_checkpoint (self, blob_names, epoch)
 
def check_db_exists (self, epoch)
 
def save (self, epoch)
 
def write_checkpoint_metadata (self, epoch)
 
def get_resume_from_epoch_id (self, user_epoch=None)
 
def set_params (self, nodes, path_prefix=None, path_type=None)
 
def cp_accessible (self, epoch=None)
 

Detailed Description

Controls saving and loading of workspaces on every epoch boundary of a job.
If a CheckpointManager instance is passed to JobRunner, then JobRunner will
call `init`, `read` and `save` at different moments in between epoch runs.

Args:
    db_prefix: The prefix used to construct full db name. Since `absolute_path`
        is set to True, this will be used as db_name in SaveOp.
    node_name: Name of the node where this checkpoint_manager is used.
    db_type: Type of database to use for storing checkpoint.
    metadata_handler: An optional object capable of reading/writing
        checkpoint info in storage of choice.

Definition at line 158 of file checkpoint.py.

Member Function Documentation

def caffe2.python.checkpoint.CheckpointManager.cp_accessible (   self,
  epoch = None 
)
Returns True if Checkpoint data is accessible

Args:
    epoch: An integer. The epoch of the checkpoint. If None,
it implies we need to check if checkpoint directory is accessible

Returns:
    is_cp_accessible: A boolean. Returns True if Checkpoint data is accessible

Definition at line 356 of file checkpoint.py.

def caffe2.python.checkpoint.CheckpointManager.get_resume_from_epoch_id (   self,
  user_epoch = None 
)
Identify the epoch-id from which Job must resume

Args:
    user_epoch: An integer. Optional parameter for user to explicitly
identify the epoch-id to load checkpoint from
Retruns:
    epoch: the epoch-id to load checkpoints from
or None if no checkpoints were written

Definition at line 319 of file checkpoint.py.

def caffe2.python.checkpoint.CheckpointManager.init (   self,
  nodes = None,
  retrieve_from_epoch = None,
  path_prefix = None,
  path_type = None 
)
Build a Task that will be run once after the job's `init_group` is run.
This task will determine which blobs need to be checkpointed.
If retrieve_from_epoch is not None, then the checkpoint metadata is
retrieved from a previously saved checkpoint.

Definition at line 202 of file checkpoint.py.

def caffe2.python.checkpoint.CheckpointManager.load (   self,
  epoch,
  path_prefix = None,
  path_type = None 
)
Build a Task that will be run by JobRunner when the job is to be
resumed from a given epoch. This task will run a Load op that will
load and deserialize all relevant blobs from a persistent storage.

Definition at line 236 of file checkpoint.py.

def caffe2.python.checkpoint.CheckpointManager.load_blobs_from_checkpoint (   self,
  blob_names,
  epoch 
)
Builds a Task that loads only the necessary blobs from a checkpoint of
the given epoch. The necessary blobs are given in the blob_names
argument.

Args:
    blob_names: A list of strings. Each string is the name of a
blob.
    epoch: The checkpoint epoch to load from.

Returns:
    A Task which loads the specified blobs from the checkpoint of the
    given epoch.

Definition at line 254 of file checkpoint.py.

def caffe2.python.checkpoint.CheckpointManager.save (   self,
  epoch 
)
Build a Task that is run once after `init_group` and after each
epoch is run. This will execute a Save ops to serialize and persist
blobs present in the global workspace.

Definition at line 294 of file checkpoint.py.

def caffe2.python.checkpoint.CheckpointManager.set_params (   self,
  nodes,
  path_prefix = None,
  path_type = None 
)
Set parameters associated with CP manager

Args:
    nodes: An array of nodes where this checkpoint manager is running.
    path_prefix: Used to construct db name or path where checkpoint files are
stored.
    path_type: Indicate the type of path where checkpoint files are stored.

Definition at line 335 of file checkpoint.py.

def caffe2.python.checkpoint.CheckpointManager.write_checkpoint_metadata (   self,
  epoch 
)
Write metadata for checkpoint

Args:
    epoch: An integer. The epoch-id for which checkpoint metadata is
written

Definition at line 308 of file checkpoint.py.


The documentation for this class was generated from the following file: