seq2seq_util.ModelHelper Class Reference

Inheritance diagram for seq2seq_util.ModelHelper:

Public Member Functions
def	__init__ (self, init_params=True)
def	AddParam (self, name, init=None, init_value=None, trainable=True)

Public Attributes
	non_trainable_params

Detailed Description

Definition at line 14 of file seq2seq_util.py.

Member Function Documentation

AddParam()

def seq2seq_util.ModelHelper.AddParam	(		self,
			name,
			init = None,
			init_value = None,
			trainable = True
	)

Adds a parameter to the model's net and it's initializer if needed

Args:
    init: a tuple (<initialization_op_name>, <initialization_op_kwargs>)
    init_value: int, float or str. Can be used instead of `init` as a
simple constant initializer
    trainable: bool, whether to compute gradient for this param or not

Definition at line 23 of file seq2seq_util.py.

---

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

• caffe2/python/examples/seq2seq_util.py