tf.contrib.framework.assign_from_checkpoint(
model_path,
var_list,
ignore_missing_vars=False
)
Defined in tensorflow/contrib/framework/python/ops/variables.py
.
Creates an operation to assign specific variables from a checkpoint.
Args:
model_path
: The full path to the model checkpoint. To get latest checkpoint usemodel_path = tf.train.latest_checkpoint(checkpoint_dir)
var_list
: A list of (possibly partitioned)Variable
objects or a dictionary mapping names in the checkpoint to the corresponding variables or list of variables to initialize from that checkpoint value. For partitioned Variables, the name in the checkpoint must be the full variable, not the name of the partitioned variable, eg. "my_var" rather than "my_var/part_4". If empty, returns no_op(), {}.ignore_missing_vars
: Boolean, if True ignore variables missing in the checkpoint with a warning instead of failing.
Returns:
the restore_op and the feed_dict that need to be run to restore var_list.
Raises:
ValueError
: Ifignore_missing_vars
is False and the checkpoint specified atmodel_path
is missing one of the variables invar_list
.