-
Notifications
You must be signed in to change notification settings - Fork 0
/
ddl_mnist_aztk.py
227 lines (178 loc) · 8.8 KB
/
ddl_mnist_aztk.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
'''
// Copyright (c) 2016 Joeri Hermans
// Copyright (c) Microsoft Corporation. All rights reserved.
// Licensed under the GPLv3 license.
MNIST classification using Distributed Keras on an AZTK Spark cluster. This script is adapted from the
dist-keras's example https://github.com/cerndb/dist-keras/blob/master/examples/mnist.py, written by Joeri
Hermans (joeri.hermans@cern.ch). To run the script, you can first create a folder named 'scripts' under
the AZTK source code folder and place it to the created folder. Then, you can submit a Spark job by
executing the following command
aztk spark cluster submit --id <my_cluster_id> --name <my_job_id> scripts/ddl_mnist.py
This example requires the following packages
- distkeras
- keras
- tensorflow
'''
import os
import time
import pyspark
from pyspark.sql import SparkSession
from distkeras.evaluators import *
from distkeras.predictors import *
from distkeras.trainers import *
from distkeras.transformers import *
from distkeras.utils import *
from keras.layers.convolutional import *
from keras.layers.core import *
from keras.models import Sequential
from keras.optimizers import *
from pyspark.ml.evaluation import MulticlassClassificationEvaluator
from pyspark.ml.feature import OneHotEncoder
from pyspark.ml.feature import StandardScaler
from pyspark.ml.feature import StringIndexer
from pyspark.ml.feature import VectorAssembler
## Setup the pyspark environment
# First, setup the Spark variables. You can modify them according to your need, e.g. increasing num_executors to reduce training time.
application_name = 'Distributed Keras MNIST'
master = 'spark://' + os.environ.get('AZTK_MASTER_IP') + ':7077'
num_processes = 2
num_executors = 4 #4, 3, 2, 1
# This variable is derived from the number of cores and executors, and will be used to assign the number of model trainers.
num_workers = num_executors * num_processes
print('Number of desired executors: ', num_executors)
print('Number of desired processes per executor: ', num_processes)
print('Total number of workers: ', num_workers)
# Use the DataBricks CSV reader, this has some nice functionality regarding invalid values.
os.environ['PYSPARK_SUBMIT_ARGS'] = '--packages com.databricks:spark-csv_2.10:1.4.0 pyspark-shell'
# Modify the Spark configuration
conf = pyspark.SparkConf()
conf.set('spark.app.name', application_name)
conf.set('spark.master', master)
conf.set('spark.executor.cores', num_processes)
conf.set('spark.executor.instances', num_executors)
conf.set('spark.executor.memory', '2g')
conf.set('spark.locality.wait', '0')
conf.set('spark.serializer', 'org.apache.spark.serializer.KryoSerializer')
conf.set('spark.local.dir', '/tmp/' + get_os_username() + '/dist-keras')
# Create the Spark context
sc = pyspark.SparkContext(conf=conf)
sqlc = pyspark.sql.SQLContext(sc)
## Load data
# Define variables
storage_account = 'publicdat'
storage_key = 'GNDru9kic+4jXle1roq4klFzwQ3Jy9CvsvMJl8o5b8W/DoX6v14PE8aPuX48V3r9yc3tuKP3NKCSUHqFobib4A=='
input_container = 'mnist'
path_train = 'wasb://{}@{}.blob.core.windows.net/mnist_train.csv'.format(input_container, storage_account)
path_test = 'wasb://{}@{}.blob.core.windows.net/mnist_test.csv'.format(input_container, storage_account)
# Attach the blob storage to the spark cluster
def attach_storage_container(spark, account, key):
config = spark._sc._jsc.hadoopConfiguration()
setting = 'fs.azure.account.key.' + account + '.blob.core.windows.net'
if not config.get(setting):
config.set(setting, key)
spark = SparkSession.builder.getOrCreate()
spark.sparkContext.setLogLevel('ERROR')
attach_storage_container(spark, storage_account, storage_key)
time.sleep(4)
# Read the training dataset
raw_dataset_train = sqlc.read.format('com.databricks.spark.csv') \
.options(header='true', inferSchema='true') \
.load(path_train)
# Read the testing dataset
raw_dataset_test = sqlc.read.format('com.databricks.spark.csv') \
.options(header='true', inferSchema='true') \
.load(path_test)
## Data transformation
# First, we extract the desired features from the raw dataset by constructing a list with all desired columns. This is identical for the test set.
features = raw_dataset_train.columns
features.remove('label')
# Next, we use Spark's VectorAssembler to assemble/create a vector of all the desired features.
# http://spark.apache.org/docs/latest/ml-features.html#vectorassembler
vector_assembler = VectorAssembler(inputCols=features, outputCol='features')
# This transformer will take all columns specified in features, and create an additional column
# 'features' which will contain all the desired features aggregated into a single vector.
dataset_train = vector_assembler.transform(raw_dataset_train)
dataset_test = vector_assembler.transform(raw_dataset_test)
# Define the number of output classes
nb_classes = 10
encoder = OneHotTransformer(nb_classes, input_col='label', output_col='label_encoded')
dataset_train = encoder.transform(dataset_train)
dataset_test = encoder.transform(dataset_test)
# Allocate a MinMaxTransformer from Distributed Keras to normalize the features
# o_min -> original_minimum
# n_min -> new_minimum
transformer = MinMaxTransformer(n_min=0.0, n_max=1.0, \
o_min=0.0, o_max=250.0, \
input_col='features', \
output_col='features_normalized')
# Transform the dataset
dataset_train = transformer.transform(dataset_train)
dataset_test = transformer.transform(dataset_test)
# Keras expects the vectors to be in a particular shape, we can reshape the vectors using Spark.
reshape_transformer = ReshapeTransformer('features_normalized', 'matrix', (28, 28, 1))
dataset_train = reshape_transformer.transform(dataset_train)
dataset_test = reshape_transformer.transform(dataset_test)
# Select the desired columns, this will reduce network usage.
dataset_train = dataset_train.select('features_normalized', 'matrix', 'label', 'label_encoded')
dataset_test = dataset_test.select('features_normalized', 'matrix', 'label', 'label_encoded')
# Keras expects DenseVectors
dense_transformer = DenseTransformer(input_col='features_normalized', output_col='features_normalized_dense')
dataset_train = dense_transformer.transform(dataset_train)
dataset_test = dense_transformer.transform(dataset_test)
# Repartition the training and test set
training_set = dataset_train.repartition(num_workers)
test_set = dataset_test.repartition(num_workers)
# Cache the data sets
training_set.cache()
test_set.cache()
# Precache the trainingset on the nodes using a simple count.
#print(training_set.count())
## Model and evaluation metric definition
# dimensions of the image
img_rows, img_cols = 28, 28
# number of convolutional filters to use
nb_filters = 32
# size of pooling area for max pooling
pool_size = (2, 2)
# convolution kernel size
kernel_size = (3, 3)
input_shape = (img_rows, img_cols, 1)
# Construct the model
convnet = Sequential()
convnet.add(Conv2D(nb_filters, (kernel_size[0], kernel_size[1]),
padding='valid',
input_shape=input_shape))
convnet.add(Activation('relu'))
convnet.add(Conv2D(nb_filters, (kernel_size[0], kernel_size[1])))
convnet.add(Activation('relu'))
convnet.add(MaxPooling2D(pool_size=pool_size))
convnet.add(Flatten())
convnet.add(Dense(225))
convnet.add(Activation('relu'))
convnet.add(Dense(nb_classes))
convnet.add(Activation('softmax'))
# Define the optimizer and the loss
optimizer_convnet = 'adam'
loss_convnet = 'categorical_crossentropy'
# Print the model structure
convnet.summary()
# Specify a procedure to evaluate the dataset in a distributed manner
def evaluate_accuracy(model, test_set, features='matrix'):
evaluator = AccuracyEvaluator(prediction_col='prediction_index', label_col='label')
predictor = ModelPredictor(keras_model=model, features_col=features)
transformer = LabelIndexTransformer(output_dim=nb_classes)
test_set = test_set.select(features, 'label')
test_set = predictor.predict(test_set)
test_set = transformer.transform(test_set)
score = evaluator.evaluate(test_set)
return score
## Model training and evaluation
# Use the ADAG optimizer. You can also use a SingleWorker to compare the performance with traditional non-distributed gradient descent.
trainer = ADAG(keras_model=convnet, worker_optimizer=optimizer_convnet, loss=loss_convnet,
num_workers=num_workers, batch_size=16, communication_window=5, num_epoch=5,
features_col='matrix', label_col='label_encoded')
trained_model = trainer.train(training_set)
print('----------------------------------------------------')
print('Training time: ', trainer.get_training_time())
print('Accuracy: ', evaluate_accuracy(trained_model, test_set))
print('Number of parameter server updates: ', trainer.parameter_server.num_updates)