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
|
import argparse
import math
import warnings
from keras import backend
from keras.layers import regularizers
from keras.regularizers import l1_l2, l1, l2
from dl.GenerateDatasetsFromSSI import vectorize_segment_info
from dl.TrainGenotypeSSIDataset import create_model, create_callbacks
from goby.SequenceSegmentInformation import SequenceSegmentInformationGenerator, SequenceSegmentInformationStreamGenerator
from goby.pyjavaproperties import Properties
import numpy as np
import tensorflow as tf
def vectorize(segment_info_generator, max_base_count, max_feature_count, max_label_count, padding="pre"):
feature_arrays = []
label_arrays = []
for segment_info in segment_info_generator:
feature_array, label_array = vectorize_segment_info(segment_info, max_base_count, max_feature_count,
max_label_count, padding)
feature_arrays.append(feature_array)
label_arrays.append(label_array)
return np.array(feature_arrays), np.array(label_arrays)
def vectorize_generator(segment_info_generator, max_base_count, max_feature_count, max_label_count, mini_batch_size,
num_segments, padding="pre"):
feature_arrays = []
label_arrays = []
segments_processed = 0
for segment_info in segment_info_generator:
segments_processed += 1
feature_array, label_array = vectorize_segment_info(segment_info, max_base_count, max_feature_count,
max_label_count, padding)
feature_arrays.append(feature_array)
label_arrays.append(label_array)
if len(feature_arrays) == mini_batch_size or segments_processed == num_segments:
yield np.array(feature_arrays), np.array(label_arrays)
feature_arrays = []
label_arrays = []
segments_processed = 0
def main(args):
backend.set_learning_phase(1)
init = tf.global_variables_initializer()
# Show device placement:
session = (tf.InteractiveSession(config=tf.ConfigProto(log_device_placement=args.show_mappings))
if args.tensorboard
else tf.Session(config=tf.ConfigProto(log_device_placement=args.show_mappings)))
session.run(init)
if args.platform == "cpu":
implementation = 0
cpu_or_gpu = "/cpu:0"
elif args.platform == "gpu":
implementation = 2
gpu_device = args.gpu_device
cpu_or_gpu = "/gpu:{}".format(gpu_device)
else:
raise ValueError("Platform {} not recognized".format(args.platform))
reg = None
if args.l1 is not None and args.l2 is not None:
reg = regularizers.get(l1_l2(args.l1, args.l2))
elif args.l1 is not None:
reg = regularizers.get(l1(args.l1))
elif args.l2 is not None:
reg = regularizers.get(l2(args.l2))
with open("{}p".format(args.input), "r") as input_ssip:
input_properties = Properties()
input_properties.load(input_ssip)
with open("{}p".format(args.validation), "r") as val_ssip:
val_properties = Properties()
val_properties.load(val_ssip)
input_base_count = int(input_properties.getProperty("maxNumOfBases"))
input_feature_count = int(input_properties.getProperty("maxNumOfFeatures"))
input_label_count = int(input_properties.getProperty("maxNumOfLabels"))
val_feature_count = int(val_properties.getProperty("maxNumOfFeatures"))
val_label_count = int(val_properties.getProperty("maxNumOfLabels"))
input_num_segments = int(input_properties.getProperty("numSegments"))
val_num_segments = int(val_properties.getProperty("numSegments"))
input_mini_batch_size = args.input_mini_batch_size
val_mini_batch_size = args.val_mini_batch_size
max_base_count = input_base_count
if input_feature_count != val_feature_count:
warnings.warn("Mismatch between input feature count {} and val feature count {}".format(input_feature_count,
val_feature_count))
if input_label_count != val_label_count:
warnings.warn("Mismatch between input label count {} and val label count {}".format(input_label_count,
val_label_count))
max_feature_count = max(input_feature_count, val_feature_count)
max_label_count = max(input_label_count, val_label_count)
print("Creating model and callbacks...")
model = create_model(num_layers=args.num_layers,
max_base_count=max_base_count,
max_feature_count=max_feature_count,
max_label_count=max_label_count,
use_bidirectional=args.bidirectional,
lstm_units=args.lstm_units,
implementation=implementation,
layer_type=args.layer_type,
learning_rate=args.learning_rate,
add_metadata=False,
regularizer=reg)
callbacks = create_callbacks(args.model_prefix, args.min_delta, args.tensorboard)
generator_training_modes = frozenset(["batch", "sequence", "batch-np", "sequence-np"])
if args.training_mode == "whole":
print("Vectorizing training data...")
training_input, training_label = vectorize(SequenceSegmentInformationGenerator(args.input),
max_base_count=max_base_count,
max_feature_count=max_feature_count,
max_label_count=max_label_count,
padding=args.padding)
print("Vectorizing validation data...")
val_input, val_label = vectorize(SequenceSegmentInformationGenerator(args.validation),
max_base_count=max_base_count,
max_feature_count=max_feature_count,
max_label_count=max_label_count,
padding=args.padding)
with tf.device(cpu_or_gpu):
print("Training...")
model.fit(x=training_input,
y=training_label,
validation_data=(val_input, val_label),
batch_size=args.mini_batch_size,
verbose=args.verbosity,
shuffle=True,
epochs=args.max_epochs,
callbacks=callbacks)
elif args.training_mode in generator_training_modes:
if args.training_mode == "batch":
input_generator = vectorize_generator(SequenceSegmentInformationStreamGenerator(args.input),
max_base_count=max_base_count,
max_feature_count=max_feature_count,
max_label_count=max_label_count,
mini_batch_size=input_mini_batch_size,
num_segments=input_num_segments,
padding=args.padding)
val_generator = vectorize_generator(SequenceSegmentInformationStreamGenerator(args.validation),
max_base_count=max_base_count,
max_feature_count=max_feature_count,
max_label_count=max_label_count,
mini_batch_size=val_mini_batch_size,
num_segments=val_num_segments,
padding=args.padding)
elif args.training_mode == "sequence":
raise Exception("sequence mode not supported yet")
else:
raise Exception("Unrecognized training mode")
input_updates = math.ceil(input_num_segments / input_mini_batch_size)
val_updates = math.ceil(val_num_segments / val_mini_batch_size)
use_multiprocessing = args.parallel is not None
num_workers = args.parallel if args.parallel is not None else 1
with tf.device(cpu_or_gpu):
print("Training...")
model.fit_generator(generator=input_generator,
steps_per_epoch=input_updates,
validation_data=val_generator,
validation_steps=val_updates,
epochs=args.max_epochs,
callbacks=callbacks,
verbose=args.verbosity,
use_multiprocessing=use_multiprocessing,
workers=num_workers)
else:
raise Exception("Unrecognized training mode")
if __name__ == "__main__":
parser = argparse.ArgumentParser()
parser.add_argument("-t", "--training-mode", type=str,
choices=["whole", "batch", "sequence"], required=True,
help="Training mode- whole loads training and validation tensors into memory at once; "
"batch creates training and validation tensors of size mini-batch-size using a generator; "
"sequence behaves similarly to batch, but uses a keras.utils.Sequence object for "
"multiprocessing")
parser.add_argument("-i", "--input", type=str, required=True,
help="SSI file with input segments.")
parser.add_argument("-v", "--validation", type=str, required=True,
help="SSI file with validation segments.")
parser.add_argument("--bidirectional", dest="bidirectional", action="store_true",
help="When set, train a bidirectional LSTM.")
parser.add_argument("--lstm-units", type=int, default=64, help="Number of LSTM units.")
parser.add_argument("--num-layers", type=int, default=1, help="Number of hidden LSTM layers.")
parser.add_argument("--platform", required=True, type=str, choices=["cpu", "gpu"],
help="Platform to train on: cpu or gpu.")
parser.add_argument("--gpu-device", type=int, default=0, help="Index of the GPU to use, when platform is gpu. "
"Not recommended, set CUDA_VISIBLE_DEVICES instead.")
parser.add_argument("--layer-type", type=str, choices=["LSTM", "RNN", "GRU", "SRU"], default="LSTM",
help="Type of RNN layer to use.")
parser.add_argument("--learning-rate", type=float, default=0.01, help="Learning rate.")
parser.add_argument("--model-prefix", type=str, default="model",
help="Prefix (a short string) to name model checkpoints with")
parser.add_argument("--min-delta", type=float, default=0, help="Minimum delta for loss improvement in each epoch.")
parser.add_argument("--tensorboard", dest="tensorboard", action="store_true",
help="When set, use an interactive session and monitor on tensorboard.")
parser.add_argument("--show-mappings", dest="show_mappings", action="store_true",
help="When set, show operation placements on cpu/gpu devices.")
parser.add_argument("--verbosity", type=int, default=1, choices=[0, 1, 2],
help="Level of verbosity, 0, not verbose, 1, progress bar, 2, one line per epoch.")
parser.add_argument("--input-mini-batch-size", type=int, default=128,
help="Number of sequences to train on at a time. Larger values increase speed, "
"but require more memory.")
parser.add_argument("--validation-mini-batch-size", type=int, default=128,
help="Number of sequences to validate on at a time. Larger values increase speed, "
"but require more memory.")
parser.add_argument("--max-epochs", type=int, default=60, help="Maximum number of epochs to train for.")
parser.add_argument("--l1", type=float, help="L1 regularization rate.")
parser.add_argument("--l2", type=float, help="L2 regularization rate.")
parser.add_argument("--parallel", type=int, help="Run training in parallel, with n workers.")
parser.add_argument("--padding", type=str, choices=["pre", "post"], default="post",
help="Whether to pad timesteps before or after sequences. Only used for whole, batch, and "
"sequence training modes.")
parser_args = parser.parse_args()
main(parser_args)
|
