cifar10_training Namespace Reference

Module caffe2.experiments.python.cifar10_training. More...

Functions
def	AddInput (model, batch_size, db, db_type)
def	AddAccuracy (model, softmax, label)
def	AddTrainingOperators (model, softmax, label, nn_model)
def	AddBookkeepingOperators (model)
def	AlexNet (model, data, args)
def	AlexNet_Prune (model, data, args)
def	ConvBNReLUDrop (model, currentblob, outputblob, input_dim, output_dim, drop_ratio=None)
def	VGG (model, data, args)
def	ResNet110 (model, data, args)
def	ResNet20 (model, data, args)
def	sparse_transform (model)
def	test_sparse (test_model)
def	trainNtest (model_gen, args)

Variables
dictionary	MODEL_TYPE_FUNCTIONS
	parser = pyinit.FbcodeArgumentParser(description='cifar-10 Tutorial')
	type
	str
	default
	choices
	help
	float
	required
	args = parser.parse_args()

Detailed Description

Module caffe2.experiments.python.cifar10_training.

Function Documentation

AddAccuracy()

def cifar10_training.AddAccuracy	(		model,
			softmax,
			label
	)

Adds an accuracy op to the model

Definition at line 35 of file cifar10_training.py.

AddBookkeepingOperators()

def cifar10_training.AddBookkeepingOperators

(

model

)

This adds a few bookkeeping operators that we can inspect later.

These operators do not affect the training procedure: they only collect
statistics and prints them to file or to logs.

Definition at line 71 of file cifar10_training.py.

AddInput()

def cifar10_training.AddInput	(		model,
			batch_size,
			db,
			db_type
	)

Adds the data input part.

Definition at line 17 of file cifar10_training.py.

AddTrainingOperators()

def cifar10_training.AddTrainingOperators	(		model,
			softmax,
			label,
			nn_model
	)

Adds training operators to the model.

Definition at line 41 of file cifar10_training.py.

ResNet110()

def cifar10_training.ResNet110	(		model,
			data,
			args
	)

Residual net as described in section 4.2 of He at. al. (2015)

Definition at line 324 of file cifar10_training.py.

ResNet20()

def cifar10_training.ResNet20	(		model,
			data,
			args
	)

Residual net as described in section 4.2 of He at. al. (2015)

Definition at line 337 of file cifar10_training.py.

VGG()

def cifar10_training.VGG	(		model,
			data,
			args
	)

Adds the VGG-Like kaggle winner Model on Cifar-10
  The original blog about the model can be found on:
      http://torch.ch/blog/2015/07/30/cifar.html

Definition at line 282 of file cifar10_training.py.

Variable Documentation

MODEL_TYPE_FUNCTIONS

dictionary cifar10_training.MODEL_TYPE_FUNCTIONS

Initial value:

=  {
    'AlexNet': AlexNet,
    'AlexNet_Prune': AlexNet_Prune,
    'VGG': VGG,
    'ResNet-110': ResNet110,
    'ResNet-20': ResNet20
}

Definition at line 467 of file cifar10_training.py.