Note

Go to the end to download the full example code.

DenseNet Classifier: Detecting Regions of Interest in Synthetic Signals#

This example demonstrates how to use DeepPeak’s DenseNet classifier to identify regions of interest (ROIs) in synthetic 1D signals containing Gaussian peaks.

We will: - Generate a dataset of noisy signals with random Gaussian peaks - Build and train a DenseNet classifier to detect ROIs - Visualize the training process and model predictions

Note

This example is fully reproducible and suitable for Sphinx-Gallery documentation.

Imports and reproducibility#

import numpy as np

from DeepPeak.machine_learning.classifier import WaveNet, BinaryIoU
from DeepPeak.signals import SignalDatasetGenerator
from DeepPeak import kernel
import DeepPeak

np.random.seed(42)

Generate synthetic dataset#

NUM_PEAKS = 3
SEQUENCE_LENGTH = 200

kernel = DeepPeak.kernel.Gaussian(
    amplitude=(10, 20),
    position=(0.1, 0.9),
    width=(0.04 * 1, 0.05 * 1),
)

generator = SignalDatasetGenerator(n_samples=1000, sequence_length=SEQUENCE_LENGTH)

dataset = generator.generate(
    kernel=kernel,
    n_peaks=(1, NUM_PEAKS),
    noise_std=0.03,
    categorical_peak_count=False,
    compute_region_of_interest=True,
)

Visualize a few example signals and their regions of interest#

_ = dataset.plot(number_of_samples=6, number_of_columns=3)
Predicted ROI (Sample 0), Predicted ROI (Sample 1), Predicted ROI (Sample 2), Predicted ROI (Sample 3), Predicted ROI (Sample 4), Predicted ROI (Sample 5)

Build and summarize the WaveNet classifier#

wavenet = WaveNet(
    sequence_length=SEQUENCE_LENGTH,
    num_filters=64,
    num_dilation_layers=3,
    kernel_size=4,
    optimizer="adam",
    loss="binary_crossentropy",
    metrics=["accuracy", BinaryIoU(threshold=0.5)],
)

wavenet.build()

<Functional name=WaveNetDetector, built=True>

Train the classifier#

history = wavenet.fit(
    dataset.signals,
    dataset.region_of_interest,
    validation_split=0.2,
    epochs=40,
    batch_size=64,
)

Epoch 1/40

 1/13 ━━━━━━━━━━━━━━━━━━━━ 26s 2s/step - BinaryIoU: 0.0000e+00 - accuracy: 0.6492 - loss: 0.7138
 2/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.0333 - accuracy: 0.6662 - loss: 0.8902   
 4/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.0686 - accuracy: 0.7175 - loss: 0.8890
 6/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.0770 - accuracy: 0.7573 - loss: 0.8783
 8/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.0792 - accuracy: 0.7851 - loss: 0.8642
10/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.0791 - accuracy: 0.8053 - loss: 0.8426
12/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.0847 - accuracy: 0.8197 - loss: 0.8177
Epoch 1: val_loss improved from None to 0.25263, saving model to /tmp/wavenet_ckpt_qrcqia5q/best.weights.h5

13/13 ━━━━━━━━━━━━━━━━━━━━ 3s 84ms/step - BinaryIoU: 0.1327 - accuracy: 0.8937 - loss: 0.6652 - val_BinaryIoU: 0.0000e+00 - val_accuracy: 0.9501 - val_loss: 0.2526
Epoch 2/40

 1/13 ━━━━━━━━━━━━━━━━━━━━ 3s 262ms/step - BinaryIoU: 0.0000e+00 - accuracy: 0.9538 - loss: 0.2471
 2/13 ━━━━━━━━━━━━━━━━━━━━ 0s 55ms/step - BinaryIoU: 0.0000e+00 - accuracy: 0.9528 - loss: 0.2528 
 3/13 ━━━━━━━━━━━━━━━━━━━━ 0s 57ms/step - BinaryIoU: 0.0000e+00 - accuracy: 0.9526 - loss: 0.2522
 4/13 ━━━━━━━━━━━━━━━━━━━━ 0s 56ms/step - BinaryIoU: 0.0000e+00 - accuracy: 0.9525 - loss: 0.2478
 5/13 ━━━━━━━━━━━━━━━━━━━━ 0s 55ms/step - BinaryIoU: 0.0077 - accuracy: 0.9527 - loss: 0.2412    
 6/13 ━━━━━━━━━━━━━━━━━━━━ 0s 55ms/step - BinaryIoU: 0.0329 - accuracy: 0.9529 - loss: 0.2345
 7/13 ━━━━━━━━━━━━━━━━━━━━ 0s 55ms/step - BinaryIoU: 0.0599 - accuracy: 0.9531 - loss: 0.2281
 8/13 ━━━━━━━━━━━━━━━━━━━━ 0s 54ms/step - BinaryIoU: 0.0860 - accuracy: 0.9534 - loss: 0.2220
 9/13 ━━━━━━━━━━━━━━━━━━━━ 0s 54ms/step - BinaryIoU: 0.1048 - accuracy: 0.9536 - loss: 0.2161
10/13 ━━━━━━━━━━━━━━━━━━━━ 0s 54ms/step - BinaryIoU: 0.1177 - accuracy: 0.9537 - loss: 0.2108
11/13 ━━━━━━━━━━━━━━━━━━━━ 0s 54ms/step - BinaryIoU: 0.1276 - accuracy: 0.9539 - loss: 0.2058
12/13 ━━━━━━━━━━━━━━━━━━━━ 0s 54ms/step - BinaryIoU: 0.1386 - accuracy: 0.9541 - loss: 0.2011
Epoch 2: val_loss improved from 0.25263 to 0.09274, saving model to /tmp/wavenet_ckpt_qrcqia5q/best.weights.h5

13/13 ━━━━━━━━━━━━━━━━━━━━ 1s 68ms/step - BinaryIoU: 0.2732 - accuracy: 0.9568 - loss: 0.1471 - val_BinaryIoU: 0.5018 - val_accuracy: 0.9592 - val_loss: 0.0927
Epoch 3/40

 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 65ms/step - BinaryIoU: 0.5278 - accuracy: 0.9635 - loss: 0.0805
 2/13 ━━━━━━━━━━━━━━━━━━━━ 0s 55ms/step - BinaryIoU: 0.5317 - accuracy: 0.9654 - loss: 0.0792
 3/13 ━━━━━━━━━━━━━━━━━━━━ 0s 54ms/step - BinaryIoU: 0.4988 - accuracy: 0.9646 - loss: 0.0815
 4/13 ━━━━━━━━━━━━━━━━━━━━ 0s 54ms/step - BinaryIoU: 0.4801 - accuracy: 0.9646 - loss: 0.0818
 5/13 ━━━━━━━━━━━━━━━━━━━━ 0s 54ms/step - BinaryIoU: 0.4753 - accuracy: 0.9650 - loss: 0.0816
 6/13 ━━━━━━━━━━━━━━━━━━━━ 0s 54ms/step - BinaryIoU: 0.4753 - accuracy: 0.9654 - loss: 0.0811
 7/13 ━━━━━━━━━━━━━━━━━━━━ 0s 53ms/step - BinaryIoU: 0.4774 - accuracy: 0.9657 - loss: 0.0807
 8/13 ━━━━━━━━━━━━━━━━━━━━ 0s 53ms/step - BinaryIoU: 0.4806 - accuracy: 0.9661 - loss: 0.0802
 9/13 ━━━━━━━━━━━━━━━━━━━━ 0s 53ms/step - BinaryIoU: 0.4830 - accuracy: 0.9665 - loss: 0.0798
10/13 ━━━━━━━━━━━━━━━━━━━━ 0s 53ms/step - BinaryIoU: 0.4848 - accuracy: 0.9669 - loss: 0.0793
11/13 ━━━━━━━━━━━━━━━━━━━━ 0s 54ms/step - BinaryIoU: 0.4869 - accuracy: 0.9672 - loss: 0.0789
12/13 ━━━━━━━━━━━━━━━━━━━━ 0s 53ms/step - BinaryIoU: 0.4891 - accuracy: 0.9675 - loss: 0.0785
Epoch 3: val_loss improved from 0.09274 to 0.07162, saving model to /tmp/wavenet_ckpt_qrcqia5q/best.weights.h5

13/13 ━━━━━━━━━━━━━━━━━━━━ 1s 65ms/step - BinaryIoU: 0.5174 - accuracy: 0.9709 - loss: 0.0732 - val_BinaryIoU: 0.5991 - val_accuracy: 0.9758 - val_loss: 0.0716
Epoch 4/40

 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 65ms/step - BinaryIoU: 0.6032 - accuracy: 0.9767 - loss: 0.0679
 2/13 ━━━━━━━━━━━━━━━━━━━━ 0s 54ms/step - BinaryIoU: 0.5944 - accuracy: 0.9772 - loss: 0.0659
 3/13 ━━━━━━━━━━━━━━━━━━━━ 0s 53ms/step - BinaryIoU: 0.5833 - accuracy: 0.9769 - loss: 0.0667
 4/13 ━━━━━━━━━━━━━━━━━━━━ 0s 53ms/step - BinaryIoU: 0.5828 - accuracy: 0.9770 - loss: 0.0663
 5/13 ━━━━━━━━━━━━━━━━━━━━ 0s 54ms/step - BinaryIoU: 0.5834 - accuracy: 0.9770 - loss: 0.0660
 6/13 ━━━━━━━━━━━━━━━━━━━━ 0s 53ms/step - BinaryIoU: 0.5831 - accuracy: 0.9769 - loss: 0.0658
 7/13 ━━━━━━━━━━━━━━━━━━━━ 0s 53ms/step - BinaryIoU: 0.5838 - accuracy: 0.9768 - loss: 0.0656
 8/13 ━━━━━━━━━━━━━━━━━━━━ 0s 53ms/step - BinaryIoU: 0.5832 - accuracy: 0.9768 - loss: 0.0654
 9/13 ━━━━━━━━━━━━━━━━━━━━ 0s 53ms/step - BinaryIoU: 0.5831 - accuracy: 0.9767 - loss: 0.0652
10/13 ━━━━━━━━━━━━━━━━━━━━ 0s 53ms/step - BinaryIoU: 0.5833 - accuracy: 0.9767 - loss: 0.0650
11/13 ━━━━━━━━━━━━━━━━━━━━ 0s 52ms/step - BinaryIoU: 0.5839 - accuracy: 0.9767 - loss: 0.0648
12/13 ━━━━━━━━━━━━━━━━━━━━ 0s 52ms/step - BinaryIoU: 0.5847 - accuracy: 0.9767 - loss: 0.0647
Epoch 4: val_loss improved from 0.07162 to 0.06727, saving model to /tmp/wavenet_ckpt_qrcqia5q/best.weights.h5

13/13 ━━━━━━━━━━━━━━━━━━━━ 1s 61ms/step - BinaryIoU: 0.5951 - accuracy: 0.9768 - loss: 0.0631 - val_BinaryIoU: 0.5818 - val_accuracy: 0.9762 - val_loss: 0.0673
Epoch 5/40

 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 61ms/step - BinaryIoU: 0.5611 - accuracy: 0.9725 - loss: 0.0749
 3/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.5874 - accuracy: 0.9747 - loss: 0.0681
 5/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.5950 - accuracy: 0.9755 - loss: 0.0652
 6/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.5963 - accuracy: 0.9758 - loss: 0.0643
 7/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.5971 - accuracy: 0.9760 - loss: 0.0638
 8/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.5981 - accuracy: 0.9761 - loss: 0.0634
 9/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.5994 - accuracy: 0.9763 - loss: 0.0630
10/13 ━━━━━━━━━━━━━━━━━━━━ 0s 51ms/step - BinaryIoU: 0.6006 - accuracy: 0.9764 - loss: 0.0627
12/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.6009 - accuracy: 0.9765 - loss: 0.0622
Epoch 5: val_loss improved from 0.06727 to 0.06357, saving model to /tmp/wavenet_ckpt_qrcqia5q/best.weights.h5

13/13 ━━━━━━━━━━━━━━━━━━━━ 1s 60ms/step - BinaryIoU: 0.6020 - accuracy: 0.9768 - loss: 0.0601 - val_BinaryIoU: 0.6160 - val_accuracy: 0.9756 - val_loss: 0.0636
Epoch 6/40

 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 61ms/step - BinaryIoU: 0.6324 - accuracy: 0.9771 - loss: 0.0603
 3/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.5973 - accuracy: 0.9766 - loss: 0.0602
 5/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.5911 - accuracy: 0.9761 - loss: 0.0608
 7/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.5938 - accuracy: 0.9761 - loss: 0.0605
 9/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.5953 - accuracy: 0.9764 - loss: 0.0600
11/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.5974 - accuracy: 0.9766 - loss: 0.0595
13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 48ms/step - BinaryIoU: 0.5991 - accuracy: 0.9767 - loss: 0.0592
Epoch 6: val_loss improved from 0.06357 to 0.05993, saving model to /tmp/wavenet_ckpt_qrcqia5q/best.weights.h5

13/13 ━━━━━━━━━━━━━━━━━━━━ 1s 59ms/step - BinaryIoU: 0.6096 - accuracy: 0.9775 - loss: 0.0574 - val_BinaryIoU: 0.6204 - val_accuracy: 0.9777 - val_loss: 0.0599
Epoch 7/40

 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 61ms/step - BinaryIoU: 0.6740 - accuracy: 0.9827 - loss: 0.0507
 3/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.6602 - accuracy: 0.9816 - loss: 0.0519
 5/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.6557 - accuracy: 0.9811 - loss: 0.0523
 6/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.6540 - accuracy: 0.9809 - loss: 0.0526
 7/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.6521 - accuracy: 0.9807 - loss: 0.0529
 8/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.6509 - accuracy: 0.9806 - loss: 0.0531
10/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.6493 - accuracy: 0.9804 - loss: 0.0532
12/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.6489 - accuracy: 0.9804 - loss: 0.0533
Epoch 7: val_loss improved from 0.05993 to 0.05771, saving model to /tmp/wavenet_ckpt_qrcqia5q/best.weights.h5

13/13 ━━━━━━━━━━━━━━━━━━━━ 1s 59ms/step - BinaryIoU: 0.6428 - accuracy: 0.9797 - loss: 0.0539 - val_BinaryIoU: 0.6525 - val_accuracy: 0.9785 - val_loss: 0.0577
Epoch 8/40

 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 62ms/step - BinaryIoU: 0.6843 - accuracy: 0.9813 - loss: 0.0527
 2/13 ━━━━━━━━━━━━━━━━━━━━ 0s 52ms/step - BinaryIoU: 0.6768 - accuracy: 0.9811 - loss: 0.0520
 4/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.6624 - accuracy: 0.9806 - loss: 0.0522
 6/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.6595 - accuracy: 0.9805 - loss: 0.0520
 8/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.6573 - accuracy: 0.9804 - loss: 0.0518
10/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.6544 - accuracy: 0.9802 - loss: 0.0520
11/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.6532 - accuracy: 0.9801 - loss: 0.0520
13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 48ms/step - BinaryIoU: 0.6517 - accuracy: 0.9801 - loss: 0.0521
Epoch 8: val_loss improved from 0.05771 to 0.05496, saving model to /tmp/wavenet_ckpt_qrcqia5q/best.weights.h5

13/13 ━━━━━━━━━━━━━━━━━━━━ 1s 59ms/step - BinaryIoU: 0.6433 - accuracy: 0.9797 - loss: 0.0524 - val_BinaryIoU: 0.6505 - val_accuracy: 0.9795 - val_loss: 0.0550
Epoch 9/40

 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 61ms/step - BinaryIoU: 0.6608 - accuracy: 0.9803 - loss: 0.0528
 3/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.6779 - accuracy: 0.9813 - loss: 0.0513
 5/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.6767 - accuracy: 0.9814 - loss: 0.0508
 7/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.6735 - accuracy: 0.9813 - loss: 0.0506
 9/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.6701 - accuracy: 0.9811 - loss: 0.0507
11/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.6687 - accuracy: 0.9811 - loss: 0.0506
13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - BinaryIoU: 0.6675 - accuracy: 0.9811 - loss: 0.0506
Epoch 9: val_loss did not improve from 0.05496

13/13 ━━━━━━━━━━━━━━━━━━━━ 1s 55ms/step - BinaryIoU: 0.6604 - accuracy: 0.9809 - loss: 0.0505 - val_BinaryIoU: 0.6600 - val_accuracy: 0.9778 - val_loss: 0.0564
Epoch 10/40

 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 61ms/step - BinaryIoU: 0.6600 - accuracy: 0.9787 - loss: 0.0537
 3/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.6579 - accuracy: 0.9789 - loss: 0.0542
 5/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.6589 - accuracy: 0.9795 - loss: 0.0532
 6/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.6590 - accuracy: 0.9796 - loss: 0.0529
 7/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.6586 - accuracy: 0.9797 - loss: 0.0527
 8/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.6587 - accuracy: 0.9798 - loss: 0.0525
 9/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.6591 - accuracy: 0.9799 - loss: 0.0523
11/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.6573 - accuracy: 0.9799 - loss: 0.0522
12/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.6569 - accuracy: 0.9799 - loss: 0.0522
Epoch 10: val_loss did not improve from 0.05496

13/13 ━━━━━━━━━━━━━━━━━━━━ 1s 56ms/step - BinaryIoU: 0.6552 - accuracy: 0.9801 - loss: 0.0512 - val_BinaryIoU: 0.5315 - val_accuracy: 0.9751 - val_loss: 0.0613
Epoch 11/40

 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 60ms/step - BinaryIoU: 0.5766 - accuracy: 0.9782 - loss: 0.0534
 3/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.6134 - accuracy: 0.9790 - loss: 0.0517
 4/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.6198 - accuracy: 0.9788 - loss: 0.0521
 5/13 ━━━━━━━━━━━━━━━━━━━━ 0s 51ms/step - BinaryIoU: 0.6212 - accuracy: 0.9788 - loss: 0.0523
 6/13 ━━━━━━━━━━━━━━━━━━━━ 0s 51ms/step - BinaryIoU: 0.6223 - accuracy: 0.9788 - loss: 0.0525
 7/13 ━━━━━━━━━━━━━━━━━━━━ 0s 51ms/step - BinaryIoU: 0.6240 - accuracy: 0.9788 - loss: 0.0526
 9/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.6281 - accuracy: 0.9788 - loss: 0.0525
11/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.6299 - accuracy: 0.9789 - loss: 0.0524
13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 48ms/step - BinaryIoU: 0.6320 - accuracy: 0.9790 - loss: 0.0522
Epoch 11: val_loss improved from 0.05496 to 0.05377, saving model to /tmp/wavenet_ckpt_qrcqia5q/best.weights.h5

13/13 ━━━━━━━━━━━━━━━━━━━━ 1s 59ms/step - BinaryIoU: 0.6439 - accuracy: 0.9795 - loss: 0.0513 - val_BinaryIoU: 0.6311 - val_accuracy: 0.9799 - val_loss: 0.0538
Epoch 12/40

 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 61ms/step - BinaryIoU: 0.5911 - accuracy: 0.9777 - loss: 0.0554
 2/13 ━━━━━━━━━━━━━━━━━━━━ 0s 51ms/step - BinaryIoU: 0.6115 - accuracy: 0.9792 - loss: 0.0529
 4/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.6345 - accuracy: 0.9805 - loss: 0.0505
 6/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.6443 - accuracy: 0.9808 - loss: 0.0497
 8/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.6490 - accuracy: 0.9809 - loss: 0.0495
10/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.6535 - accuracy: 0.9810 - loss: 0.0492
12/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.6576 - accuracy: 0.9812 - loss: 0.0489
Epoch 12: val_loss improved from 0.05377 to 0.05047, saving model to /tmp/wavenet_ckpt_qrcqia5q/best.weights.h5

13/13 ━━━━━━━━━━━━━━━━━━━━ 1s 59ms/step - BinaryIoU: 0.6803 - accuracy: 0.9822 - loss: 0.0468 - val_BinaryIoU: 0.6789 - val_accuracy: 0.9806 - val_loss: 0.0505
Epoch 13/40

 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 62ms/step - BinaryIoU: 0.6563 - accuracy: 0.9793 - loss: 0.0517
 3/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.6666 - accuracy: 0.9798 - loss: 0.0508
 5/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.6683 - accuracy: 0.9804 - loss: 0.0496
 6/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.6705 - accuracy: 0.9807 - loss: 0.0490
 7/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.6724 - accuracy: 0.9809 - loss: 0.0485
 9/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.6761 - accuracy: 0.9813 - loss: 0.0478
11/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.6781 - accuracy: 0.9815 - loss: 0.0475
13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 48ms/step - BinaryIoU: 0.6793 - accuracy: 0.9816 - loss: 0.0472
Epoch 13: val_loss improved from 0.05047 to 0.04892, saving model to /tmp/wavenet_ckpt_qrcqia5q/best.weights.h5

13/13 ━━━━━━━━━━━━━━━━━━━━ 1s 59ms/step - BinaryIoU: 0.6853 - accuracy: 0.9823 - loss: 0.0461 - val_BinaryIoU: 0.6889 - val_accuracy: 0.9811 - val_loss: 0.0489
Epoch 14/40

 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 60ms/step - BinaryIoU: 0.7045 - accuracy: 0.9827 - loss: 0.0439
 3/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.6888 - accuracy: 0.9815 - loss: 0.0449
 5/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.6878 - accuracy: 0.9816 - loss: 0.0449
 7/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.6888 - accuracy: 0.9818 - loss: 0.0446
 9/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.6889 - accuracy: 0.9819 - loss: 0.0445
11/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.6888 - accuracy: 0.9821 - loss: 0.0444
13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - BinaryIoU: 0.6896 - accuracy: 0.9821 - loss: 0.0443
Epoch 14: val_loss improved from 0.04892 to 0.04685, saving model to /tmp/wavenet_ckpt_qrcqia5q/best.weights.h5

13/13 ━━━━━━━━━━━━━━━━━━━━ 1s 58ms/step - BinaryIoU: 0.6928 - accuracy: 0.9826 - loss: 0.0442 - val_BinaryIoU: 0.6919 - val_accuracy: 0.9815 - val_loss: 0.0468
Epoch 15/40

 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 61ms/step - BinaryIoU: 0.7238 - accuracy: 0.9837 - loss: 0.0416
 3/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.7171 - accuracy: 0.9833 - loss: 0.0425
 5/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.7106 - accuracy: 0.9832 - loss: 0.0427
 7/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.7064 - accuracy: 0.9830 - loss: 0.0429
 9/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.7030 - accuracy: 0.9828 - loss: 0.0432
11/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.7019 - accuracy: 0.9828 - loss: 0.0432
13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - BinaryIoU: 0.7017 - accuracy: 0.9829 - loss: 0.0431
Epoch 15: val_loss did not improve from 0.04685

13/13 ━━━━━━━━━━━━━━━━━━━━ 1s 55ms/step - BinaryIoU: 0.7007 - accuracy: 0.9832 - loss: 0.0427 - val_BinaryIoU: 0.6641 - val_accuracy: 0.9803 - val_loss: 0.0469
Epoch 16/40

 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 61ms/step - BinaryIoU: 0.6889 - accuracy: 0.9836 - loss: 0.0380
 2/13 ━━━━━━━━━━━━━━━━━━━━ 0s 51ms/step - BinaryIoU: 0.6893 - accuracy: 0.9832 - loss: 0.0390
 3/13 ━━━━━━━━━━━━━━━━━━━━ 0s 51ms/step - BinaryIoU: 0.6864 - accuracy: 0.9827 - loss: 0.0405
 5/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.6827 - accuracy: 0.9820 - loss: 0.0424
 7/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.6832 - accuracy: 0.9818 - loss: 0.0431
 9/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.6832 - accuracy: 0.9818 - loss: 0.0434
11/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.6841 - accuracy: 0.9818 - loss: 0.0435
12/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.6844 - accuracy: 0.9819 - loss: 0.0434
Epoch 16: val_loss did not improve from 0.04685

13/13 ━━━━━━━━━━━━━━━━━━━━ 1s 56ms/step - BinaryIoU: 0.6850 - accuracy: 0.9821 - loss: 0.0430 - val_BinaryIoU: 0.6804 - val_accuracy: 0.9806 - val_loss: 0.0489
Epoch 17/40

 1/13 ━━━━━━━━━━━━━━━━━━━━ 7s 589ms/step - BinaryIoU: 0.6973 - accuracy: 0.9827 - loss: 0.0424
 3/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.6913 - accuracy: 0.9825 - loss: 0.0425 
 5/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.6919 - accuracy: 0.9826 - loss: 0.0423
 6/13 ━━━━━━━━━━━━━━━━━━━━ 0s 51ms/step - BinaryIoU: 0.6929 - accuracy: 0.9827 - loss: 0.0422
 8/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.6950 - accuracy: 0.9828 - loss: 0.0419
10/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.6942 - accuracy: 0.9827 - loss: 0.0420
12/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.6924 - accuracy: 0.9826 - loss: 0.0422
Epoch 17: val_loss did not improve from 0.04685

13/13 ━━━━━━━━━━━━━━━━━━━━ 1s 55ms/step - BinaryIoU: 0.6839 - accuracy: 0.9822 - loss: 0.0427 - val_BinaryIoU: 0.6402 - val_accuracy: 0.9734 - val_loss: 0.0578
Epoch 18/40

 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 61ms/step - BinaryIoU: 0.6219 - accuracy: 0.9736 - loss: 0.0553
 2/13 ━━━━━━━━━━━━━━━━━━━━ 0s 53ms/step - BinaryIoU: 0.6356 - accuracy: 0.9758 - loss: 0.0526
 3/13 ━━━━━━━━━━━━━━━━━━━━ 0s 52ms/step - BinaryIoU: 0.6429 - accuracy: 0.9771 - loss: 0.0515
 4/13 ━━━━━━━━━━━━━━━━━━━━ 0s 51ms/step - BinaryIoU: 0.6511 - accuracy: 0.9781 - loss: 0.0503
 6/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.6604 - accuracy: 0.9791 - loss: 0.0489
 8/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.6652 - accuracy: 0.9797 - loss: 0.0479
10/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.6687 - accuracy: 0.9801 - loss: 0.0472
11/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.6697 - accuracy: 0.9802 - loss: 0.0470
12/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.6709 - accuracy: 0.9803 - loss: 0.0467
Epoch 18: val_loss improved from 0.04685 to 0.04467, saving model to /tmp/wavenet_ckpt_qrcqia5q/best.weights.h5

13/13 ━━━━━━━━━━━━━━━━━━━━ 1s 59ms/step - BinaryIoU: 0.6869 - accuracy: 0.9821 - loss: 0.0432 - val_BinaryIoU: 0.6928 - val_accuracy: 0.9819 - val_loss: 0.0447
Epoch 19/40

 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 61ms/step - BinaryIoU: 0.7405 - accuracy: 0.9855 - loss: 0.0362
 2/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.7162 - accuracy: 0.9841 - loss: 0.0395
 4/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.7059 - accuracy: 0.9833 - loss: 0.0413
 6/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.6991 - accuracy: 0.9829 - loss: 0.0418
 7/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.6982 - accuracy: 0.9828 - loss: 0.0418
 8/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.6973 - accuracy: 0.9827 - loss: 0.0419
 9/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.6966 - accuracy: 0.9827 - loss: 0.0420
11/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.6953 - accuracy: 0.9826 - loss: 0.0420
13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 48ms/step - BinaryIoU: 0.6949 - accuracy: 0.9826 - loss: 0.0419
Epoch 19: val_loss improved from 0.04467 to 0.04289, saving model to /tmp/wavenet_ckpt_qrcqia5q/best.weights.h5

13/13 ━━━━━━━━━━━━━━━━━━━━ 1s 59ms/step - BinaryIoU: 0.6927 - accuracy: 0.9827 - loss: 0.0414 - val_BinaryIoU: 0.7068 - val_accuracy: 0.9831 - val_loss: 0.0429
Epoch 20/40

 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 61ms/step - BinaryIoU: 0.7684 - accuracy: 0.9884 - loss: 0.0333
 3/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.7393 - accuracy: 0.9857 - loss: 0.0366
 5/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.7332 - accuracy: 0.9854 - loss: 0.0368
 7/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.7306 - accuracy: 0.9853 - loss: 0.0369
 9/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.7290 - accuracy: 0.9852 - loss: 0.0370
11/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.7284 - accuracy: 0.9851 - loss: 0.0371
12/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.7278 - accuracy: 0.9851 - loss: 0.0372
Epoch 20: val_loss improved from 0.04289 to 0.04167, saving model to /tmp/wavenet_ckpt_qrcqia5q/best.weights.h5

13/13 ━━━━━━━━━━━━━━━━━━━━ 1s 59ms/step - BinaryIoU: 0.7216 - accuracy: 0.9845 - loss: 0.0381 - val_BinaryIoU: 0.7125 - val_accuracy: 0.9835 - val_loss: 0.0417
Epoch 21/40

 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 61ms/step - BinaryIoU: 0.7463 - accuracy: 0.9854 - loss: 0.0352
 2/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.7256 - accuracy: 0.9846 - loss: 0.0373
 3/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.7242 - accuracy: 0.9847 - loss: 0.0374
 4/13 ━━━━━━━━━━━━━━━━━━━━ 0s 51ms/step - BinaryIoU: 0.7249 - accuracy: 0.9848 - loss: 0.0374
 5/13 ━━━━━━━━━━━━━━━━━━━━ 0s 51ms/step - BinaryIoU: 0.7242 - accuracy: 0.9848 - loss: 0.0374
 6/13 ━━━━━━━━━━━━━━━━━━━━ 0s 51ms/step - BinaryIoU: 0.7229 - accuracy: 0.9848 - loss: 0.0374
 8/13 ━━━━━━━━━━━━━━━━━━━━ 0s 51ms/step - BinaryIoU: 0.7205 - accuracy: 0.9846 - loss: 0.0377
10/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.7193 - accuracy: 0.9845 - loss: 0.0379
12/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.7179 - accuracy: 0.9844 - loss: 0.0382
Epoch 21: val_loss did not improve from 0.04167

13/13 ━━━━━━━━━━━━━━━━━━━━ 1s 56ms/step - BinaryIoU: 0.7115 - accuracy: 0.9838 - loss: 0.0395 - val_BinaryIoU: 0.6730 - val_accuracy: 0.9825 - val_loss: 0.0451
Epoch 22/40

 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 63ms/step - BinaryIoU: 0.6588 - accuracy: 0.9830 - loss: 0.0414
 2/13 ━━━━━━━━━━━━━━━━━━━━ 0s 51ms/step - BinaryIoU: 0.6822 - accuracy: 0.9842 - loss: 0.0385
 4/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.7011 - accuracy: 0.9846 - loss: 0.0377
 6/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.7069 - accuracy: 0.9846 - loss: 0.0378
 8/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.7087 - accuracy: 0.9845 - loss: 0.0379
10/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.7109 - accuracy: 0.9845 - loss: 0.0377
12/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.7129 - accuracy: 0.9846 - loss: 0.0376
Epoch 22: val_loss improved from 0.04167 to 0.04163, saving model to /tmp/wavenet_ckpt_qrcqia5q/best.weights.h5

13/13 ━━━━━━━━━━━━━━━━━━━━ 1s 59ms/step - BinaryIoU: 0.7211 - accuracy: 0.9846 - loss: 0.0370 - val_BinaryIoU: 0.6793 - val_accuracy: 0.9824 - val_loss: 0.0416
Epoch 23/40

 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 61ms/step - BinaryIoU: 0.6568 - accuracy: 0.9820 - loss: 0.0408
 3/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.6874 - accuracy: 0.9826 - loss: 0.0400
 5/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.6970 - accuracy: 0.9831 - loss: 0.0399
 7/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.7036 - accuracy: 0.9834 - loss: 0.0395
 8/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.7058 - accuracy: 0.9836 - loss: 0.0393
10/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.7089 - accuracy: 0.9837 - loss: 0.0391
11/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.7102 - accuracy: 0.9838 - loss: 0.0389
13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - BinaryIoU: 0.7123 - accuracy: 0.9839 - loss: 0.0387
Epoch 23: val_loss improved from 0.04163 to 0.04125, saving model to /tmp/wavenet_ckpt_qrcqia5q/best.weights.h5

13/13 ━━━━━━━━━━━━━━━━━━━━ 1s 58ms/step - BinaryIoU: 0.7247 - accuracy: 0.9848 - loss: 0.0370 - val_BinaryIoU: 0.7273 - val_accuracy: 0.9832 - val_loss: 0.0413
Epoch 24/40

 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 63ms/step - BinaryIoU: 0.7500 - accuracy: 0.9866 - loss: 0.0324
 2/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.7550 - accuracy: 0.9870 - loss: 0.0320
 4/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.7516 - accuracy: 0.9869 - loss: 0.0320
 5/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.7503 - accuracy: 0.9868 - loss: 0.0322
 7/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.7473 - accuracy: 0.9866 - loss: 0.0327
 9/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.7458 - accuracy: 0.9864 - loss: 0.0330
10/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.7453 - accuracy: 0.9863 - loss: 0.0332
12/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.7436 - accuracy: 0.9862 - loss: 0.0335
Epoch 24: val_loss did not improve from 0.04125

13/13 ━━━━━━━━━━━━━━━━━━━━ 1s 55ms/step - BinaryIoU: 0.7361 - accuracy: 0.9855 - loss: 0.0351 - val_BinaryIoU: 0.6413 - val_accuracy: 0.9809 - val_loss: 0.0470
Epoch 25/40

 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 60ms/step - BinaryIoU: 0.6822 - accuracy: 0.9840 - loss: 0.0354
 3/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.7084 - accuracy: 0.9850 - loss: 0.0340
 5/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.7110 - accuracy: 0.9847 - loss: 0.0353
 7/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.7094 - accuracy: 0.9844 - loss: 0.0366
 8/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.7092 - accuracy: 0.9843 - loss: 0.0370
 9/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.7089 - accuracy: 0.9842 - loss: 0.0372
10/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.7079 - accuracy: 0.9841 - loss: 0.0376
11/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.7073 - accuracy: 0.9840 - loss: 0.0379
12/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.7066 - accuracy: 0.9839 - loss: 0.0381
Epoch 25: val_loss did not improve from 0.04125

13/13 ━━━━━━━━━━━━━━━━━━━━ 1s 56ms/step - BinaryIoU: 0.6999 - accuracy: 0.9830 - loss: 0.0409 - val_BinaryIoU: 0.6676 - val_accuracy: 0.9823 - val_loss: 0.0433
Epoch 26/40

 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 62ms/step - BinaryIoU: 0.6624 - accuracy: 0.9813 - loss: 0.0461
 2/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.6752 - accuracy: 0.9827 - loss: 0.0432
 3/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.6836 - accuracy: 0.9831 - loss: 0.0423
 5/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.6940 - accuracy: 0.9835 - loss: 0.0408
 7/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.6992 - accuracy: 0.9837 - loss: 0.0402
 9/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.7037 - accuracy: 0.9839 - loss: 0.0396
11/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.7063 - accuracy: 0.9840 - loss: 0.0394
13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - BinaryIoU: 0.7088 - accuracy: 0.9841 - loss: 0.0391
Epoch 26: val_loss did not improve from 0.04125

13/13 ━━━━━━━━━━━━━━━━━━━━ 1s 55ms/step - BinaryIoU: 0.7237 - accuracy: 0.9848 - loss: 0.0374 - val_BinaryIoU: 0.6782 - val_accuracy: 0.9826 - val_loss: 0.0413
Epoch 27/40

 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 60ms/step - BinaryIoU: 0.7211 - accuracy: 0.9849 - loss: 0.0364
 3/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.7276 - accuracy: 0.9853 - loss: 0.0358
 5/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.7322 - accuracy: 0.9855 - loss: 0.0353
 6/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.7312 - accuracy: 0.9854 - loss: 0.0355
 7/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.7305 - accuracy: 0.9854 - loss: 0.0355
 8/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.7305 - accuracy: 0.9854 - loss: 0.0355
10/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.7309 - accuracy: 0.9854 - loss: 0.0355
12/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.7305 - accuracy: 0.9854 - loss: 0.0356
Epoch 27: val_loss improved from 0.04125 to 0.04011, saving model to /tmp/wavenet_ckpt_qrcqia5q/best.weights.h5

13/13 ━━━━━━━━━━━━━━━━━━━━ 1s 59ms/step - BinaryIoU: 0.7315 - accuracy: 0.9852 - loss: 0.0361 - val_BinaryIoU: 0.6809 - val_accuracy: 0.9826 - val_loss: 0.0401
Epoch 28/40

 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 62ms/step - BinaryIoU: 0.7061 - accuracy: 0.9848 - loss: 0.0345
 2/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.7054 - accuracy: 0.9850 - loss: 0.0340
 4/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.7144 - accuracy: 0.9853 - loss: 0.0339
 6/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.7193 - accuracy: 0.9854 - loss: 0.0340
 8/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.7205 - accuracy: 0.9853 - loss: 0.0342
10/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.7219 - accuracy: 0.9853 - loss: 0.0343
11/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.7224 - accuracy: 0.9853 - loss: 0.0343
12/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.7229 - accuracy: 0.9853 - loss: 0.0344
Epoch 28: val_loss improved from 0.04011 to 0.03680, saving model to /tmp/wavenet_ckpt_qrcqia5q/best.weights.h5

13/13 ━━━━━━━━━━━━━━━━━━━━ 1s 59ms/step - BinaryIoU: 0.7300 - accuracy: 0.9853 - loss: 0.0347 - val_BinaryIoU: 0.7292 - val_accuracy: 0.9849 - val_loss: 0.0368
Epoch 29/40

 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 61ms/step - BinaryIoU: 0.7466 - accuracy: 0.9868 - loss: 0.0302
 3/13 ━━━━━━━━━━━━━━━━━━━━ 0s 48ms/step - BinaryIoU: 0.7519 - accuracy: 0.9867 - loss: 0.0308
 5/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.7462 - accuracy: 0.9863 - loss: 0.0321
 7/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.7438 - accuracy: 0.9861 - loss: 0.0327
 9/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.7431 - accuracy: 0.9860 - loss: 0.0330
11/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.7431 - accuracy: 0.9860 - loss: 0.0332
13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - BinaryIoU: 0.7425 - accuracy: 0.9859 - loss: 0.0334
Epoch 29: val_loss did not improve from 0.03680

13/13 ━━━━━━━━━━━━━━━━━━━━ 1s 55ms/step - BinaryIoU: 0.7395 - accuracy: 0.9858 - loss: 0.0342 - val_BinaryIoU: 0.7316 - val_accuracy: 0.9826 - val_loss: 0.0423
Epoch 30/40

 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 62ms/step - BinaryIoU: 0.7545 - accuracy: 0.9851 - loss: 0.0369
 2/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.7437 - accuracy: 0.9850 - loss: 0.0359
 4/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.7377 - accuracy: 0.9850 - loss: 0.0354
 6/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.7387 - accuracy: 0.9851 - loss: 0.0350
 8/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.7396 - accuracy: 0.9852 - loss: 0.0349
10/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.7406 - accuracy: 0.9852 - loss: 0.0347
12/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.7415 - accuracy: 0.9854 - loss: 0.0345
Epoch 30: val_loss did not improve from 0.03680

13/13 ━━━━━━━━━━━━━━━━━━━━ 1s 55ms/step - BinaryIoU: 0.7488 - accuracy: 0.9864 - loss: 0.0326 - val_BinaryIoU: 0.7329 - val_accuracy: 0.9832 - val_loss: 0.0400
Epoch 31/40

 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 60ms/step - BinaryIoU: 0.7669 - accuracy: 0.9855 - loss: 0.0355
 3/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.7656 - accuracy: 0.9865 - loss: 0.0336
 5/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.7607 - accuracy: 0.9866 - loss: 0.0330
 7/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.7589 - accuracy: 0.9866 - loss: 0.0326
 8/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.7581 - accuracy: 0.9866 - loss: 0.0325
10/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.7579 - accuracy: 0.9867 - loss: 0.0323
11/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.7578 - accuracy: 0.9867 - loss: 0.0323
13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 48ms/step - BinaryIoU: 0.7577 - accuracy: 0.9867 - loss: 0.0322
Epoch 31: val_loss improved from 0.03680 to 0.03607, saving model to /tmp/wavenet_ckpt_qrcqia5q/best.weights.h5

13/13 ━━━━━━━━━━━━━━━━━━━━ 1s 59ms/step - BinaryIoU: 0.7568 - accuracy: 0.9868 - loss: 0.0320 - val_BinaryIoU: 0.7452 - val_accuracy: 0.9847 - val_loss: 0.0361
Epoch 32/40

 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 61ms/step - BinaryIoU: 0.7626 - accuracy: 0.9860 - loss: 0.0337
 3/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.7702 - accuracy: 0.9870 - loss: 0.0313
 5/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.7702 - accuracy: 0.9871 - loss: 0.0311
 7/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.7697 - accuracy: 0.9871 - loss: 0.0311
 9/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.7660 - accuracy: 0.9870 - loss: 0.0311
11/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.7636 - accuracy: 0.9869 - loss: 0.0313
12/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.7625 - accuracy: 0.9869 - loss: 0.0313
Epoch 32: val_loss did not improve from 0.03607

13/13 ━━━━━━━━━━━━━━━━━━━━ 1s 55ms/step - BinaryIoU: 0.7471 - accuracy: 0.9862 - loss: 0.0324 - val_BinaryIoU: 0.7367 - val_accuracy: 0.9835 - val_loss: 0.0397
Epoch 33/40

 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 60ms/step - BinaryIoU: 0.7488 - accuracy: 0.9834 - loss: 0.0407
 3/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.7542 - accuracy: 0.9847 - loss: 0.0377
 5/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.7511 - accuracy: 0.9849 - loss: 0.0371
 7/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.7497 - accuracy: 0.9850 - loss: 0.0367
 8/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.7501 - accuracy: 0.9852 - loss: 0.0363
 9/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.7503 - accuracy: 0.9853 - loss: 0.0360
11/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.7513 - accuracy: 0.9855 - loss: 0.0354
13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 48ms/step - BinaryIoU: 0.7526 - accuracy: 0.9857 - loss: 0.0349
Epoch 33: val_loss improved from 0.03607 to 0.03340, saving model to /tmp/wavenet_ckpt_qrcqia5q/best.weights.h5

13/13 ━━━━━━━━━━━━━━━━━━━━ 1s 59ms/step - BinaryIoU: 0.7607 - accuracy: 0.9870 - loss: 0.0316 - val_BinaryIoU: 0.7630 - val_accuracy: 0.9868 - val_loss: 0.0334
Epoch 34/40

 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 62ms/step - BinaryIoU: 0.7701 - accuracy: 0.9877 - loss: 0.0298
 2/13 ━━━━━━━━━━━━━━━━━━━━ 0s 52ms/step - BinaryIoU: 0.7642 - accuracy: 0.9871 - loss: 0.0313
 3/13 ━━━━━━━━━━━━━━━━━━━━ 0s 51ms/step - BinaryIoU: 0.7615 - accuracy: 0.9868 - loss: 0.0323
 5/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.7586 - accuracy: 0.9867 - loss: 0.0323
 7/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.7579 - accuracy: 0.9868 - loss: 0.0320
 9/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.7580 - accuracy: 0.9868 - loss: 0.0320
11/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.7581 - accuracy: 0.9868 - loss: 0.0319
12/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.7579 - accuracy: 0.9868 - loss: 0.0318
Epoch 34: val_loss did not improve from 0.03340

13/13 ━━━━━━━━━━━━━━━━━━━━ 1s 56ms/step - BinaryIoU: 0.7573 - accuracy: 0.9869 - loss: 0.0312 - val_BinaryIoU: 0.7443 - val_accuracy: 0.9840 - val_loss: 0.0379
Epoch 35/40

 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 60ms/step - BinaryIoU: 0.7951 - accuracy: 0.9882 - loss: 0.0286
 3/13 ━━━━━━━━━━━━━━━━━━━━ 0s 48ms/step - BinaryIoU: 0.7797 - accuracy: 0.9877 - loss: 0.0302
 5/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.7704 - accuracy: 0.9873 - loss: 0.0311
 6/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.7684 - accuracy: 0.9872 - loss: 0.0312
 7/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.7671 - accuracy: 0.9872 - loss: 0.0313
 8/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.7658 - accuracy: 0.9872 - loss: 0.0313
10/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.7652 - accuracy: 0.9872 - loss: 0.0312
12/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.7652 - accuracy: 0.9872 - loss: 0.0312
Epoch 35: val_loss did not improve from 0.03340

13/13 ━━━━━━━━━━━━━━━━━━━━ 1s 55ms/step - BinaryIoU: 0.7618 - accuracy: 0.9872 - loss: 0.0313 - val_BinaryIoU: 0.7341 - val_accuracy: 0.9834 - val_loss: 0.0392
Epoch 36/40

 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 61ms/step - BinaryIoU: 0.7409 - accuracy: 0.9850 - loss: 0.0334
 3/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.7453 - accuracy: 0.9859 - loss: 0.0320
 5/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.7503 - accuracy: 0.9862 - loss: 0.0314
 7/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.7562 - accuracy: 0.9866 - loss: 0.0309
 9/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.7587 - accuracy: 0.9868 - loss: 0.0307
11/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.7607 - accuracy: 0.9869 - loss: 0.0305
13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 47ms/step - BinaryIoU: 0.7625 - accuracy: 0.9871 - loss: 0.0303
Epoch 36: val_loss did not improve from 0.03340

13/13 ━━━━━━━━━━━━━━━━━━━━ 1s 55ms/step - BinaryIoU: 0.7737 - accuracy: 0.9878 - loss: 0.0290 - val_BinaryIoU: 0.7505 - val_accuracy: 0.9864 - val_loss: 0.0345
Epoch 37/40

 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 60ms/step - BinaryIoU: 0.7940 - accuracy: 0.9892 - loss: 0.0302
 3/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.7976 - accuracy: 0.9892 - loss: 0.0286
 4/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.7977 - accuracy: 0.9891 - loss: 0.0285
 6/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.7955 - accuracy: 0.9890 - loss: 0.0282
 8/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.7933 - accuracy: 0.9889 - loss: 0.0281
 9/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.7923 - accuracy: 0.9888 - loss: 0.0281
11/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.7911 - accuracy: 0.9888 - loss: 0.0280
13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 48ms/step - BinaryIoU: 0.7903 - accuracy: 0.9887 - loss: 0.0280
Epoch 37: val_loss improved from 0.03340 to 0.03166, saving model to /tmp/wavenet_ckpt_qrcqia5q/best.weights.h5

13/13 ━━━━━━━━━━━━━━━━━━━━ 1s 59ms/step - BinaryIoU: 0.7851 - accuracy: 0.9886 - loss: 0.0279 - val_BinaryIoU: 0.7737 - val_accuracy: 0.9873 - val_loss: 0.0317
Epoch 38/40

 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 60ms/step - BinaryIoU: 0.7660 - accuracy: 0.9866 - loss: 0.0301
 3/13 ━━━━━━━━━━━━━━━━━━━━ 0s 51ms/step - BinaryIoU: 0.7727 - accuracy: 0.9871 - loss: 0.0296
 5/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.7701 - accuracy: 0.9871 - loss: 0.0301
 7/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.7726 - accuracy: 0.9873 - loss: 0.0298
 9/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.7740 - accuracy: 0.9875 - loss: 0.0296
11/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.7747 - accuracy: 0.9876 - loss: 0.0295
12/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.7749 - accuracy: 0.9876 - loss: 0.0295
Epoch 38: val_loss did not improve from 0.03166

13/13 ━━━━━━━━━━━━━━━━━━━━ 1s 55ms/step - BinaryIoU: 0.7786 - accuracy: 0.9882 - loss: 0.0287 - val_BinaryIoU: 0.7716 - val_accuracy: 0.9864 - val_loss: 0.0326
Epoch 39/40

 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 60ms/step - BinaryIoU: 0.8221 - accuracy: 0.9903 - loss: 0.0241
 3/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.8073 - accuracy: 0.9894 - loss: 0.0253
 5/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.7981 - accuracy: 0.9891 - loss: 0.0260
 7/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.7907 - accuracy: 0.9886 - loss: 0.0269
 9/13 ━━━━━━━━━━━━━━━━━━━━ 0s 50ms/step - BinaryIoU: 0.7870 - accuracy: 0.9884 - loss: 0.0274
11/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.7842 - accuracy: 0.9882 - loss: 0.0278
13/13 ━━━━━━━━━━━━━━━━━━━━ 0s 48ms/step - BinaryIoU: 0.7829 - accuracy: 0.9882 - loss: 0.0280
Epoch 39: val_loss did not improve from 0.03166

13/13 ━━━━━━━━━━━━━━━━━━━━ 1s 55ms/step - BinaryIoU: 0.7761 - accuracy: 0.9879 - loss: 0.0289 - val_BinaryIoU: 0.7186 - val_accuracy: 0.9852 - val_loss: 0.0344
Epoch 40/40

 1/13 ━━━━━━━━━━━━━━━━━━━━ 0s 60ms/step - BinaryIoU: 0.7430 - accuracy: 0.9879 - loss: 0.0293
 3/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.7645 - accuracy: 0.9885 - loss: 0.0279
 4/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.7688 - accuracy: 0.9885 - loss: 0.0277
 6/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.7750 - accuracy: 0.9886 - loss: 0.0276
 8/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.7789 - accuracy: 0.9887 - loss: 0.0273
10/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.7817 - accuracy: 0.9888 - loss: 0.0271
12/13 ━━━━━━━━━━━━━━━━━━━━ 0s 49ms/step - BinaryIoU: 0.7832 - accuracy: 0.9888 - loss: 0.0271
Epoch 40: val_loss improved from 0.03166 to 0.02985, saving model to /tmp/wavenet_ckpt_qrcqia5q/best.weights.h5

13/13 ━━━━━━━━━━━━━━━━━━━━ 1s 59ms/step - BinaryIoU: 0.7907 - accuracy: 0.9890 - loss: 0.0269 - val_BinaryIoU: 0.7850 - val_accuracy: 0.9880 - val_loss: 0.0298
Restored best model weights from /tmp/wavenet_ckpt_qrcqia5q/best.weights.h5

Plot training history#

_ = wavenet.plot_model_history()
BinaryIoU, accuracy, loss, val_BinaryIoU, val_accuracy, val_loss

Predict and visualize on a test signal#

_ = wavenet.plot_prediction(dataset=dataset, number_of_samples=12, number_of_columns=3, threshold=0.1, randomize_signal=True)
Predicted ROI (Sample 632), Predicted ROI (Sample 89), Predicted ROI (Sample 369), Predicted ROI (Sample 131), Predicted ROI (Sample 354), Predicted ROI (Sample 792), Predicted ROI (Sample 797), Predicted ROI (Sample 952), Predicted ROI (Sample 831), Predicted ROI (Sample 395), Predicted ROI (Sample 693), Predicted ROI (Sample 825)

Total running time of the script: (0 minutes 38.283 seconds)

Gallery generated by Sphinx-Gallery