-
Notifications
You must be signed in to change notification settings - Fork 0
/
cnn_2d.py
168 lines (144 loc) · 5.02 KB
/
cnn_2d.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
import json
import random
import numpy as np
import tensorflow as tf
from tensorflow.keras.callbacks import ModelCheckpoint, ReduceLROnPlateau
from tensorflow.keras.layers import (
Activation,
Conv2D,
Dense,
Dropout,
Flatten,
MaxPooling2D,
)
from tensorflow.keras.models import Sequential
from tensorflow.keras.regularizers import l2
from data_utils import normalize_min_max_v2, prepare_data, shuffle_data
RAND_SEED = 1
np.random.seed(RAND_SEED)
tf.random.set_seed(RAND_SEED)
random.seed(RAND_SEED)
def train_cnn(train_set, test_set, window_size, overlap, exp_name):
X_train, y_train = prepare_data(train_set)
X_test, y_test = prepare_data(test_set)
X_train = normalize_min_max_v2(X_train, 0, 1)
X_test = normalize_min_max_v2(X_test, 0, 1)
# reshape only when grayscale
X_train = X_train.reshape(X_train.shape[0], X_train.shape[1], X_train.shape[2], 1)
X_test = X_test.reshape(X_test.shape[0], X_test.shape[1], X_test.shape[2], 1)
DROPOUT_RATE = 0.15
KERNEL_SIZE = (5, 5)
L2_REGULARIZATION = 0.0001
model = Sequential(
[
Conv2D(
20,
KERNEL_SIZE,
padding="same",
kernel_regularizer=l2(L2_REGULARIZATION),
input_shape=(X_train.shape[1], X_train.shape[2], X_train.shape[3]),
),
Activation("relu"),
MaxPooling2D(
(2, 2),
),
# Dropout(DROPOUT_RATE),
Conv2D(
40,
KERNEL_SIZE,
padding="same",
kernel_regularizer=l2(L2_REGULARIZATION),
),
Activation("relu"),
MaxPooling2D((2, 2)),
# Dropout(DROPOUT_RATE),
# Conv2D(
# 80,
# KERNEL_SIZE,
# padding="same",
# kernel_regularizer=l2(L2_REGULARIZATION),
# ),
# Activation("relu"),
# MaxPooling2D((2, 2)),
# # Dropout(DROPOUT_RATE),
Flatten(),
Dense(
512,
# kernel_regularizer=l2(L2_REGULARIZATION)
),
Activation("relu"),
Dropout(DROPOUT_RATE),
Dense(20, activation="softmax"),
]
)
model.summary()
model.compile(
optimizer="adam", loss="categorical_crossentropy", metrics=["accuracy"]
)
# Define the path where you want to save the best checkpoint
checkpoint_filepath = (
f"results/cnn2d_models/cnn2d_{window_size}_{overlap}_{exp_name}.h5"
)
# Define the ModelCheckpoint callback
model_checkpoint = ModelCheckpoint(
checkpoint_filepath,
monitor="val_accuracy", # Choose the metric to monitor (e.g., validation loss)
mode="max", # 'min' for metrics like validation loss, 'max' for accuracy, etc.
save_best_only=True, # Save only the best model checkpoint
save_weights_only=False, # Save the entire model, not just weights
verbose=1, # Print messages about checkpoint saving
)
# Learning rate scheduler
lr_scheduler = ReduceLROnPlateau(
monitor="loss", min_delta=0.01, factor=0.6, patience=1, min_lr=0.0000001
)
metrics = {
"train_losses": [],
"test_losses": [],
"train_accuracies": [],
"test_accuracies": [],
}
for epoch in range(500):
X_train, y_train = shuffle_data(X_train, y_train)
history = model.fit(
X_train,
y_train,
epochs=1,
batch_size=40,
verbose=0,
callbacks=[lr_scheduler, model_checkpoint],
validation_data=(X_test, y_test),
) # Train for one epoch at a time
# Store training metrics
metrics["train_losses"].append(history.history["loss"][0])
metrics["train_accuracies"].append(history.history["accuracy"][0])
# Store test/validation metrics
metrics["test_losses"].append(history.history["val_loss"][0])
metrics["test_accuracies"].append(history.history["val_accuracy"][0])
# save metrics in json file
with open(
f"results/cnn2d_models/cnn2d_{window_size}_{overlap}_{exp_name}.json", "w"
) as fp:
json.dump(metrics, fp)
NPY_DATA_DIR = "npy_datasets"
spectogram_map = {
256: [8, 64, 128, 250],
512: [0, 256, 511],
}
window_size = 256
overlap = 8
for window_size, overlap_list in spectogram_map.items():
for overlap in overlap_list:
# Load the dataset
test_dataset = np.load(
f"{NPY_DATA_DIR}/test_{window_size}_{overlap}_png.npy", allow_pickle=True
)
retest_dataset = np.load(
f"{NPY_DATA_DIR}/retest_{window_size}_{overlap}_png.npy", allow_pickle=True
)
test_retest_metrics = train_cnn(
test_dataset, retest_dataset, window_size, overlap, exp_name="test_retest"
)
retest_test_metrics = train_cnn(
retest_dataset, test_dataset, window_size, overlap, exp_name="retest_test"
)