caffe_translator Namespace Reference

Module caffe2.python.caffe_translator !/usr/bin/env python2. More...

Classes
class	TranslatorRegistry

Functions
def	TranslateModel (args, kwargs)
def	ConvertTensorProtosToInitNet (net_params, input_name)
def	BaseTranslate (layer, caffe2_type)
def	AddArgument (op, key, value)
def	TranslateInput (layer, pretrained_blobs, is_test)
	Common translators for layers.
def	TranslateData (layer, pretrained_blobs, is_test)
def	TranslateConv (layer, pretrained_blobs, is_test)
def	TranslateDeconv (layer, pretrained_blobs, is_test)
def	TranslateRelu (layer, pretrained_blobs, is_test)
def	TranslatePool (layer, pretrained_blobs, is_test)
def	TranslateLRN (layer, pretrained_blobs, is_test)
def	TranslateInnerProduct (layer, pretrained_blobs, is_test)
def	TranslateDropout (layer, pretrained_blobs, is_test)
def	TranslateSoftmax (layer, pretrained_blobs, is_test)
def	TranslateSoftmaxWithLoss (layer, pretrained_blobs, is_test)
def	TranslateAccuracy (layer, pretrained_blobs, is_test)
def	TranslateConcat (layer, pretrained_blobs, is_test)
def	TranslateTanH (layer, pretrained_blobs, is_test)
def	TranslateInstanceNorm (layer, pretrained_blobs, is_test)
def	TranslateElementWise (layer, pretrained_blobs, is_test)
def	TranslateScale (layer, pretrained_blobs, is_test)
def	TranslateReshape (layer, pretrained_blobs, is_test)
def	TranslateFlatten (layer, pretrained_blobs, is_test)
def	TranslateSigmoid (layer, pretrained_blobs, is_test)
def	TranslateROIPooling (layer, pretrained_blobs, is_test)
def	TranslatePRelu (layer, pretrained_blobs, is_test)
def	TranslateReduction (layer, pretrained_blobs, is_test)

Variables
	log = logging.getLogger("caffe_translator")
	parser
	help
	default
	args = parser.parse_args()
	caffenet = caffe_pb2.NetParameter()
	caffenet_pretrained = caffe_pb2.NetParameter()
	input_proto = args.prototext
	input_caffemodel = args.caffemodel
	output_init_net = args.init_net
	output_predict_net = args.predict_net
	net
	pretrained_params
	is_test
	external_input = net.op[0].input[0]
	external_output = net.op[-1].output[0]
def	init_net = ConvertTensorProtosToInitNet(pretrained_params, external_input)

Detailed Description

Module caffe2.python.caffe_translator !/usr/bin/env python2.

Function Documentation

AddArgument()

def caffe_translator.AddArgument	(		op,
			key,
			value
	)

Makes an argument based on the value type.

Definition at line 165 of file caffe_translator.py.

BaseTranslate()

def caffe_translator.BaseTranslate	(		layer,
			caffe2_type
	)

A simple translate interface that maps the layer input and output.

Definition at line 156 of file caffe_translator.py.

ConvertTensorProtosToInitNet()

def caffe_translator.ConvertTensorProtosToInitNet	(		net_params,
			input_name
	)

Takes the net_params returned from TranslateModel, and wrap it as an
init net that contain GivenTensorFill.

This is a very simple feature that only works with float tensors, and is
only intended to be used in an environment where you want a single
initialization file - for more complex cases, use a db to store the
parameters.

Definition at line 132 of file caffe_translator.py.

Variable Documentation

parser

caffe_translator.parser

Initial value:

=  argparse.ArgumentParser(
        description="Utilitity to convert pretrained caffe models to Caffe2 models.")

Definition at line 580 of file caffe_translator.py.