This repository has been archived by the owner on Jan 24, 2024. It is now read-only.
-
Notifications
You must be signed in to change notification settings - Fork 0
/
build_model.py
110 lines (81 loc) · 3.57 KB
/
build_model.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
import os
import pickle
import numpy as np
import tensorflow as tf
# Batch size should be a divisor of number of training samples
batch_size = 94
num_epochs = 22
print(f'Running on TensorFlow v.{tf.__version__} with Numpy v.{np.__version__}')
print('> ====== Loading compressed training data')
# Load processed compressed data from disk
pickle_in = open('x_train.pickle', 'rb')
x_train = pickle.load(pickle_in)
pickle_in = open('y_train.pickle', 'rb')
y_train = pickle.load(pickle_in)
pickle_in = open('x_valid.pickle', 'rb')
x_valid = pickle.load(pickle_in)
pickle_in = open('y_valid.pickle', 'rb')
y_valid = pickle.load(pickle_in)
print('> ====== Compressed training data loaded')
def define_model():
model = tf.keras.models.Sequential()
# Convolutional feature map layers
model.add(tf.keras.layers.BatchNormalization(input_shape=x_train.shape[1:]))
model.add(tf.keras.layers.Conv2D(24, (5, 5), padding='same', activation='elu'))
model.add(tf.keras.layers.MaxPooling2D(pool_size=(2, 2)))
model.add(tf.keras.layers.Dropout(0.1))
model.add(tf.keras.layers.BatchNormalization(input_shape=x_train.shape[1:]))
model.add(tf.keras.layers.Conv2D(36, (5, 5), padding='same', activation='elu'))
model.add(tf.keras.layers.MaxPooling2D(pool_size=(2, 2)))
model.add(tf.keras.layers.Dropout(0.1))
model.add(tf.keras.layers.BatchNormalization(input_shape=x_train.shape[1:]))
model.add(tf.keras.layers.Conv2D(48, (5, 5), padding='same', activation='elu'))
model.add(tf.keras.layers.MaxPooling2D(pool_size=(2, 2)))
model.add(tf.keras.layers.Dropout(0.1))
model.add(tf.keras.layers.BatchNormalization(input_shape=x_train.shape[1:]))
model.add(tf.keras.layers.Conv2D(64, (3, 3), padding='same', activation='elu'))
model.add(tf.keras.layers.MaxPooling2D(pool_size=(2, 2)))
model.add(tf.keras.layers.Dropout(0.1))
model.add(tf.keras.layers.BatchNormalization(input_shape=x_train.shape[1:]))
model.add(tf.keras.layers.Conv2D(64, (3, 3), padding='same', activation='elu'))
model.add(tf.keras.layers.MaxPooling2D(pool_size=(2, 2)))
model.add(tf.keras.layers.Dropout(0.1))
# Convert the 3D feature maps to 1D feature vectors
model.add(tf.keras.layers.Flatten())
# Fully-connected layers
model.add(tf.keras.layers.Dense(100))
model.add(tf.keras.layers.Activation('elu'))
model.add(tf.keras.layers.Dropout(0.2))
model.add(tf.keras.layers.Dense(50))
model.add(tf.keras.layers.Activation('elu'))
model.add(tf.keras.layers.Dropout(0.2))
model.add(tf.keras.layers.Dense(10))
model.add(tf.keras.layers.Activation('elu'))
model.add(tf.keras.layers.Dropout(0.2))
# Output layer: normalized steering angle
model.add(tf.keras.layers.Dense(1))
model.add(tf.keras.layers.Activation('softmax'))
return model
# Obtain a resolver and connect to a TPU runtime
resolver = tf.contrib.cluster_resolver.TPUClusterResolver('grpc://' + os.environ['COLAB_TPU_ADDR'])
tf.contrib.distribute.initialize_tpu_system(resolver)
strategy = tf.contrib.distribute.TPUStrategy(resolver)
# Define and compile the model
with strategy.scope():
model = define_model()
model.compile(
optimizer=tf.keras.optimizers.Adam(lr=1e-3),
loss='mse',
metrics=['mse', 'mae']
)
# Train the model
model.fit(
x_train.astype(np.float32),
y_train.astype(np.float32),
epochs=num_epochs,
# batch_size=batch_size,
steps_per_epoch=94,
validation_data=(x_valid.astype(np.float32), y_valid.astype(np.float32))
)
model.save_weights('jetCNN.h5', overwrite=True)
print('>>>>>>> Training complete')