tf.config.experimental.set_memory_growth() failed, err: list index out of range
Selected GPU : Quadro RTX 6000 (id=0)
Model: "model"
_________________________________________________________________
Layer (type)                 Output Shape              Param #   
=================================================================
input_1 (InputLayer)         [(None, 96, 96, 3)]       0         
_________________________________________________________________
conv2d (Conv2D)              (None, 48, 48, 8)         224       
_________________________________________________________________
batch_normalization (BatchNo (None, 48, 48, 8)         32        
_________________________________________________________________
activation (Activation)      (None, 48, 48, 8)         0         
_________________________________________________________________
depthwise_conv2d (DepthwiseC (None, 48, 48, 8)         80        
_________________________________________________________________
batch_normalization_1 (Batch (None, 48, 48, 8)         32        
_________________________________________________________________
activation_1 (Activation)    (None, 48, 48, 8)         0         
_________________________________________________________________
conv2d_1 (Conv2D)            (None, 48, 48, 16)        144       
_________________________________________________________________
batch_normalization_2 (Batch (None, 48, 48, 16)        64        
_________________________________________________________________
activation_2 (Activation)    (None, 48, 48, 16)        0         
_________________________________________________________________
depthwise_conv2d_1 (Depthwis (None, 24, 24, 16)        160       
_________________________________________________________________
batch_normalization_3 (Batch (None, 24, 24, 16)        64        
_________________________________________________________________
activation_3 (Activation)    (None, 24, 24, 16)        0         
_________________________________________________________________
conv2d_2 (Conv2D)            (None, 24, 24, 32)        544       
_________________________________________________________________
batch_normalization_4 (Batch (None, 24, 24, 32)        128       
_________________________________________________________________
activation_4 (Activation)    (None, 24, 24, 32)        0         
_________________________________________________________________
depthwise_conv2d_2 (Depthwis (None, 24, 24, 32)        320       
_________________________________________________________________
batch_normalization_5 (Batch (None, 24, 24, 32)        128       
_________________________________________________________________
activation_5 (Activation)    (None, 24, 24, 32)        0         
_________________________________________________________________
conv2d_3 (Conv2D)            (None, 24, 24, 32)        1056      
_________________________________________________________________
batch_normalization_6 (Batch (None, 24, 24, 32)        128       
_________________________________________________________________
activation_6 (Activation)    (None, 24, 24, 32)        0         
_________________________________________________________________
depthwise_conv2d_3 (Depthwis (None, 12, 12, 32)        320       
_________________________________________________________________
batch_normalization_7 (Batch (None, 12, 12, 32)        128       
_________________________________________________________________
activation_7 (Activation)    (None, 12, 12, 32)        0         
_________________________________________________________________
conv2d_4 (Conv2D)            (None, 12, 12, 64)        2112      
_________________________________________________________________
batch_normalization_8 (Batch (None, 12, 12, 64)        256       
_________________________________________________________________
activation_8 (Activation)    (None, 12, 12, 64)        0         
_________________________________________________________________
depthwise_conv2d_4 (Depthwis (None, 12, 12, 64)        640       
_________________________________________________________________
batch_normalization_9 (Batch (None, 12, 12, 64)        256       
_________________________________________________________________
activation_9 (Activation)    (None, 12, 12, 64)        0         
_________________________________________________________________
conv2d_5 (Conv2D)            (None, 12, 12, 64)        4160      
_________________________________________________________________
batch_normalization_10 (Batc (None, 12, 12, 64)        256       
_________________________________________________________________
activation_10 (Activation)   (None, 12, 12, 64)        0         
_________________________________________________________________
depthwise_conv2d_5 (Depthwis (None, 6, 6, 64)          640       
_________________________________________________________________
batch_normalization_11 (Batc (None, 6, 6, 64)          256       
_________________________________________________________________
activation_11 (Activation)   (None, 6, 6, 64)          0         
_________________________________________________________________
conv2d_6 (Conv2D)            (None, 6, 6, 128)         8320      
_________________________________________________________________
batch_normalization_12 (Batc (None, 6, 6, 128)         512       
_________________________________________________________________
activation_12 (Activation)   (None, 6, 6, 128)         0         
_________________________________________________________________
depthwise_conv2d_6 (Depthwis (None, 6, 6, 128)         1280      
_________________________________________________________________
batch_normalization_13 (Batc (None, 6, 6, 128)         512       
_________________________________________________________________
activation_13 (Activation)   (None, 6, 6, 128)         0         
_________________________________________________________________
conv2d_7 (Conv2D)            (None, 6, 6, 128)         16512     
_________________________________________________________________
batch_normalization_14 (Batc (None, 6, 6, 128)         512       
_________________________________________________________________
activation_14 (Activation)   (None, 6, 6, 128)         0         
_________________________________________________________________
depthwise_conv2d_7 (Depthwis (None, 6, 6, 128)         1280      
_________________________________________________________________
batch_normalization_15 (Batc (None, 6, 6, 128)         512       
_________________________________________________________________
activation_15 (Activation)   (None, 6, 6, 128)         0         
_________________________________________________________________
conv2d_8 (Conv2D)            (None, 6, 6, 128)         16512     
_________________________________________________________________
batch_normalization_16 (Batc (None, 6, 6, 128)         512       
_________________________________________________________________
activation_16 (Activation)   (None, 6, 6, 128)         0         
_________________________________________________________________
depthwise_conv2d_8 (Depthwis (None, 6, 6, 128)         1280      
_________________________________________________________________
batch_normalization_17 (Batc (None, 6, 6, 128)         512       
_________________________________________________________________
activation_17 (Activation)   (None, 6, 6, 128)         0         
_________________________________________________________________
conv2d_9 (Conv2D)            (None, 6, 6, 128)         16512     
_________________________________________________________________
batch_normalization_18 (Batc (None, 6, 6, 128)         512       
_________________________________________________________________
activation_18 (Activation)   (None, 6, 6, 128)         0         
_________________________________________________________________
depthwise_conv2d_9 (Depthwis (None, 6, 6, 128)         1280      
_________________________________________________________________
batch_normalization_19 (Batc (None, 6, 6, 128)         512       
_________________________________________________________________
activation_19 (Activation)   (None, 6, 6, 128)         0         
_________________________________________________________________
conv2d_10 (Conv2D)           (None, 6, 6, 128)         16512     
_________________________________________________________________
batch_normalization_20 (Batc (None, 6, 6, 128)         512       
_________________________________________________________________
activation_20 (Activation)   (None, 6, 6, 128)         0         
_________________________________________________________________
depthwise_conv2d_10 (Depthwi (None, 6, 6, 128)         1280      
_________________________________________________________________
batch_normalization_21 (Batc (None, 6, 6, 128)         512       
_________________________________________________________________
activation_21 (Activation)   (None, 6, 6, 128)         0         
_________________________________________________________________
conv2d_11 (Conv2D)           (None, 6, 6, 128)         16512     
_________________________________________________________________
batch_normalization_22 (Batc (None, 6, 6, 128)         512       
_________________________________________________________________
activation_22 (Activation)   (None, 6, 6, 128)         0         
_________________________________________________________________
depthwise_conv2d_11 (Depthwi (None, 3, 3, 128)         1280      
_________________________________________________________________
batch_normalization_23 (Batc (None, 3, 3, 128)         512       
_________________________________________________________________
activation_23 (Activation)   (None, 3, 3, 128)         0         
_________________________________________________________________
conv2d_12 (Conv2D)           (None, 3, 3, 256)         33024     
_________________________________________________________________
batch_normalization_24 (Batc (None, 3, 3, 256)         1024      
_________________________________________________________________
activation_24 (Activation)   (None, 3, 3, 256)         0         
_________________________________________________________________
depthwise_conv2d_12 (Depthwi (None, 3, 3, 256)         2560      
_________________________________________________________________
batch_normalization_25 (Batc (None, 3, 3, 256)         1024      
_________________________________________________________________
activation_25 (Activation)   (None, 3, 3, 256)         0         
_________________________________________________________________
conv2d_13 (Conv2D)           (None, 3, 3, 256)         65792     
_________________________________________________________________
batch_normalization_26 (Batc (None, 3, 3, 256)         1024      
_________________________________________________________________
activation_26 (Activation)   (None, 3, 3, 256)         0         
_________________________________________________________________
average_pooling2d (AveragePo (None, 1, 1, 256)         0         
_________________________________________________________________
flatten (Flatten)            (None, 256)               0         
_________________________________________________________________
dense (Dense)                (None, 2)                 514       
=================================================================
Total params: 221,794
Trainable params: 216,322
Non-trainable params: 5,472
_________________________________________________________________

Total MACs: 7.490 M
Total OPs: 15.677 M
Name: visual_wake_words
Version: 1
Description: TinyML: Visual Wake Words - MobileNetv1 with COCO14
Classes: person, non_person
hash: None
date: None
Training dataset: Found 98657 samples belonging to 2 classes:
    person = 47826
non_person = 50831
Validation dataset: Found 10962 samples belonging to 2 classes:
    person = 5314
non_person = 5648
Using default TensorBoard callback with following parameters:
{'histogram_freq': 1,
 'log_dir': '/home/dariedle/.mltk/models/visual_wake_words/train/tensorboard',
 'profile_batch': 2,
 'update_freq': 'epoch',
 'write_graph': True,
 'write_images': False}
Using default ModelCheckpoint callback with following parameters:
{'filepath': '/home/dariedle/.mltk/models/visual_wake_words/train/weights/weights-{epoch:03d}-{val_accuracy:.4f}.h5',
 'mode': 'auto',
 'monitor': 'val_accuracy',
 'options': None,
 'save_best_only': True,
 'save_freq': 'epoch',
 'save_weights_only': True,
 'verbose': 0}
Using default LearningRateScheduler callback with following parameters:
{'schedule': <function lr_schedule at 0x7f516c2cdf70>, 'verbose': 1}
Enabling model checkpoints
Using Keras callbacks: TensorBoard, ModelCheckpoint, LearningRateScheduler, ModelCheckpoint
Starting model training ...
Generating /home/dariedle/.mltk/models/visual_wake_words/visual_wake_words.h5


*** Best training val_accuracy = 0.852


Generating /home/dariedle/.mltk/models/visual_wake_words/train/training-history.json
Generating /home/dariedle/.mltk/models/visual_wake_words/train/training-history.png
Creating /data/dariedle/mltk/mltk/models/tinyml/visual_wake_words.mltk.zip
