def extract_contour_dataset(dataset_location='classification/sample_data/imgs_recycling/',
dataset_name='recycling',
enable_logging=True):
# if we're not running inside AML WB, set up the share directory
if 'AZUREML_NATIVE_SHARE_DIRECTORY' not in os.environ:
os.environ['AZUREML_NATIVE_SHARE_DIRECTORY'] = './share'
# create a dataset from a directory with folders representing different classes
dataset = ClassificationDataset.create_from_dir(dataset_name,
dataset_location,
enable_logging=enable_logging)
extract_contour(dataset)
def classify(dataset_location='classification/sample_data/imgs_recycling/',
dataset_name='recycling',
do_augmentations=False):
"""
a sample pipeline for classification.
loads a dataset, optionally does some augmentations, creates and trains a
model using transfer learning based on ResNet18, and returns the accuracy
on a test set.
Args:
dataset_location: path to a dataset. there should be a top level folder
containing folders for each class. see the sample recycling dataset for
an example of the format
dataset_name: the of the dataset. will be used in the dataset
management functionality
do_augmentations: boolean. specifies whether augmentations should be
applied to the test set
Returns:
the accuracy on the test set
"""
# if we're not running inside AML WB, set up the share directory
if 'AZUREML_NATIVE_SHARE_DIRECTORY' not in os.environ:
os.environ['AZUREML_NATIVE_SHARE_DIRECTORY'] = './share'
# create a dataset from a directory with folders representing different classes
dataset = ClassificationDataset.create_from_dir(dataset_name,
dataset_location)
# print out some info about the dataset
dataset.print_info()
# split the full dataset into a train and test set
# the stratify option will ensure that the different labels are balanced in the
# train and test sets
splitter = Splitter(dataset)
train_set_orig, test_set = splitter.split(train_size=.8, stratify='label')
# optionally augment images by cropping and rotating
if do_augmentations:
# here we create two pipelines for doing augmentations. the first
# will rotate each image by between -45 and 45 degrees (the angle is
# chosen at random). then the rotated images will be flipped from left
# to right with probability .5. the second pipeline will randomly crop
# images by between 0 and 10 percent. each pipeline will be applied to
# the original dataset. the resulting dataset will three times as many
# images as the original - the original dataset, the dataset after
# augmentation by the rotate_and_flip pipeline, and the dataset
# after augmentation by the crop pipeline
rotate_and_flip = augmenters.Sequential(
[augmenters.Affine(rotate=(-45, 45)),
augmenters.Fliplr(.5)])
crop = augmenters.Sequential([augmenters.Crop(percent=(0, .1))])
train_set = augment_dataset(train_set_orig, [rotate_and_flip, crop])
else:
train_set = train_set_orig
# now create the model
base_model_name = 'ResNet18_ImageNet_CNTK'
model = CNTKTLModel(train_set.labels,
base_model_name=base_model_name,
output_path='.')
# train the model using cntk
model.train(train_set)
# return the accuracy
ce = ClassificationEvaluation(model, test_set, minibatch_size=16)
acc = ce.compute_accuracy()
return acc
def train_deploy(dataset_location='classification/sample_data/imgs_recycling',
dataset_name='recycling', do_augmentations=True,
deployment_name="testdeployment", azureml_rscgroup=None, azureml_cluster_name=None):
"""
a sample pipeline for deploying themodel that is trained on a dataset.
loads a dataset, optionally does some augmentations, creates and trains a
model using transfer learning based on ResNet18, deploys the trained model
on the specified Azure ML cluster or picks up the one set using the CLI.
and returns the Scoring URL.
Args:
dataset_location: path to a dataset. there should be a top level folder
containing folders for each class. see the sample recycling dataset for
an example of the format
dataset_name: the of the dataset. will be used in the dataset
management functionality
do_augmentations: boolean. specifies whether augmentations should be
applied to the test set
deployment_name: the deployment of the deployment. Will be used in deployment
management facility
azureml_rscgroup: Azure ML resource group name of the model management account.
If not set, default value will be picked up if set from CLI
azureml_cluster_name: Azure ML cluster name where the model is deployed. If not set,
default value will be picked up if set from CLI.
Returns:
the scoring API URL of the deployment
"""
# if we're not running inside AML WB, set up the share directory
if 'AZUREML_NATIVE_SHARE_DIRECTORY' not in os.environ:
os.environ['AZUREML_NATIVE_SHARE_DIRECTORY'] = './share'
context = Context.get_global_context()
# create a dataset from a directory with folders representing different classes
dataset = ClassificationDataset.create_from_dir(dataset_name, dataset_location)
# print out some info about the dataset
dataset.print_info()
# optionally augment images by cropping and rotating
if do_augmentations:
# here we create two pipelines for doing augmentations. the first
# will rotate each image by between -45 and 45 degrees (the angle is
# chosen at random). then the rotated images will be flipped from left
# to right with probability .5. the second pipeline will randomly crop
# images by between 0 and 10 percent. each pipeline will be applied to
# the original dataset. the resulting dataset will three times as many
# images as the original - the original dataset, the dataset after
# augmentation by the rotate_and_flip pipeline, and the dataset
# after augmentation by the crop pipeline
rotate_and_flip = augmenters.Sequential([
augmenters.Affine(rotate=(-45, 45)),
augmenters.Fliplr(.5)])
crop = augmenters.Sequential([augmenters.Crop(percent=(0, .1))])
train_set = augment_dataset(dataset, [rotate_and_flip, crop])
else:
train_set = dataset
# now create the model
base_model_name = 'ResNet18_ImageNet_CNTK'
model = CNTKTLModel(train_set.labels,
base_model_name = base_model_name,
output_path='.')
# train the model using cntk
num_epochs = 5
mb_size = 32
model.train(train_set,
lr_per_mb=[.01] * 20 + [.001] * 20 + [.0001],
num_epochs=num_epochs,
mb_size=mb_size)
print("Model state:", model.model_state)
# check if the deployment exists, if yes remove it first
AMLDeployment.delete_if_service_exist(deployment_name)
#deploy the trained model
deploy_obj = AMLDeployment(
deployment_name=deployment_name, associated_DNNModel=model, aml_env = "cluster", replicas=1)
deploy_obj.deploy()
return deploy_obj.service_url
# Dataset Creation
if 'AZUREML_NATIVE_SHARE_DIRECTORY' not in os.environ:
os.environ['AZUREML_NATIVE_SHARE_DIRECTORY'] = './share'
context = Context.get_global_context()
dataset_name = "fashion"
dataset_location = os.path.join(
Context.get_global_context().storage.outputs_path, "data",
dataset_name)
print("Dataset Location:", dataset_location)
print("Downloading images to: " + dataset_location)
download_images.download_all(dataset_location)
dataset = ClassificationDataset.create_from_dir(dataset_name,
dataset_location)
print("Dataset consists of {} images with {} labels.".format(
len(dataset.images), len(dataset.labels)))
# Split the data into train and test
splitter = Splitter(dataset)
train_set, test_set = splitter.split(train_size=.5,
random_state=1,
stratify="label")
print("Number of original training images = {}.".format(train_set.size()))
num_train_sets = 20
num_test_sets = 20
num_different_label = 50
trainPairs = ImagePairs(train_set, num_train_sets, num_different_label)
print(
'There are {} sets of image pairs generated for all labels from training data.'
def classify(dataset_location='classification/sample_data/imgs_recycling/',
dataset_name='recycling', do_augmentations=True,
enable_logging=True):
"""
a sample pipeline for classification.
loads a dataset, optionally does some augmentations, creates and trains a
model using transfer learning based on ResNet18, and returns the accuracy
on a test set.
Args:
dataset_location: path to a dataset. there should be a top level folder
containing folders for each class. see the sample recycling dataset for
an example of the format
dataset_name: the of the dataset. will be used in the dataset
management functionality
do_augmentations: boolean. specifies whether augmentations should be
applied to the test set
Returns:
the accuracy on the test set
"""
# if we're not running inside AML WB, set up the share directory
if 'AZUREML_NATIVE_SHARE_DIRECTORY' not in os.environ:
os.environ['AZUREML_NATIVE_SHARE_DIRECTORY'] = './share'
# create a dataset from a directory with folders representing different classes
dataset = ClassificationDataset.create_from_dir(dataset_name,
dataset_location)
# print out some info about the dataset
print("DATASET INFO:")
dataset.print_info()
# split the full dataset into a train and test set
# the stratify option will ensure that the different labels are balanced in the
# train and test sets
train_set_orig, test_set = dataset.split(train_size = 0.66, stratify = "label")
# optionally augment images by cropping and rotating
if do_augmentations:
aug_sequence = augmenters.Sequential([
augmenters.Fliplr(0.5), # horizontally flip 50% of all images
augmenters.Crop(percent=(0, 0.1)) # crop images by 0-10% of their height/width
])
train_set = augment_dataset(train_set_orig, [aug_sequence])
print("Number of original training images = {}, with augmented images included = {}.".format(train_set_orig.size(), train_set.size()))
else:
train_set = train_set_orig
# model creation
lr_per_mb = [0.05]*7 + [0.005]*7 + [0.0005]
mb_size = 32
input_resoluton = 224
base_model_name = 'ResNet18_ImageNet_CNTK'
model = CNTKTLModel(train_set.labels,
base_model_name=base_model_name,
image_dims = (3, input_resoluton, input_resoluton))
# train the model using cntk
ce = ClassificationEvaluation(model, test_set, minibatch_size = mb_size)
acc = ce.compute_accuracy()
print("Accuracy = {:2.2f}%".format(100*acc))
cm = ce.compute_confusion_matrix()
print("Confusion matrix = \n{}".format(cm))
cm_ui = ConfusionMatrixUI(cm, [l.name for l in test_set.labels])
show(cm_ui.ui)
return acc
def create_dataset_from_json():
file_labels = "C:\\Users\\miprasad\\Downloads\\cvp-1.0.0b2-release5\\cvp-1.0.0b2-release\\cvp_project\\classification\\scripts\\file_labels.json"
dataset = ClassificationDataset.create_from_json("recycling", file_labels, context=None)
dataset.print_info()