caffe2.python.predictor.predictor_exporter.PredictorExportMeta Class Reference

Inheritance diagram for caffe2.python.predictor.predictor_exporter.PredictorExportMeta:

Public Member Functions
def	__new__ (cls, predict_net, parameters, inputs, outputs, shapes=None, name="", extra_init_net=None, net_type=None, num_workers=None, trainer_prefix=None)
def	inputs_name (self)
def	outputs_name (self)
def	parameters_name (self)
def	global_init_name (self)
def	predict_init_name (self)
def	predict_net_name (self)
def	train_init_plan_name (self)
def	train_plan_name (self)

Detailed Description

Metadata to be used for serializaing a net.

parameters, inputs, outputs could be either BlobReference or blob's names

predict_net can be either core.Net, NetDef, PlanDef or object

Override the named tuple to provide optional name parameter.
name will be used to identify multiple prediction nets.

net_type is the type field in caffe2 NetDef - can be 'simple', 'dag', etc.

num_workers specifies for net type 'dag' how many threads should run ops

trainer_prefix specifies the type of trainer.

Definition at line 39 of file predictor_exporter.py.

---

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

• caffe2/python/predictor/predictor_exporter.py