'''
filters = input._keras_shape[-1] # channel_axis = -1 for TF
se = GlobalAveragePooling1D()(input)
se = Reshape((1, filters))(se)
se = Dense(filters // 16,
activation='relu',
kernel_initializer='he_normal',
use_bias=False)(se)
se = Dense(filters,
activation='sigmoid',
kernel_initializer='he_normal',
use_bias=False)(se)
se = multiply([input, se])
return se
if __name__ == "__main__":
model = generate_model_2()
train_model(model,
DATASET_INDEX,
dataset_prefix='eeg2_attention',
epochs=500,
batch_size=128)
evaluate_model(model,
DATASET_INDEX,
dataset_prefix='eeg2_attention',
batch_size=128)
'''
filters = input._keras_shape[-1] # channel_axis = -1 for TF
se = GlobalAveragePooling1D()(input)
se = Reshape((1, filters))(se)
se = Dense(filters // 16,
activation='relu',
kernel_initializer='he_normal',
use_bias=False)(se)
se = Dense(filters,
activation='sigmoid',
kernel_initializer='he_normal',
use_bias=False)(se)
se = multiply([input, se])
return se
if __name__ == "__main__":
model = generate_model_2()
train_model(model,
DATASET_INDEX,
dataset_prefix='arabic_voice_',
epochs=1000,
batch_size=128)
evaluate_model(model,
DATASET_INDEX,
dataset_prefix='arabic_voice_',
batch_size=128)
filters: number of output filters
k: width factor
Returns: a keras tensor
'''
filters = input._keras_shape[-1] # channel_axis = -1 for TF
se = GlobalAveragePooling1D()(input)
se = Reshape((1, filters))(se)
se = Dense(filters // 16,
activation='relu',
kernel_initializer='he_normal',
use_bias=False)(se)
se = Dense(filters,
activation='sigmoid',
kernel_initializer='he_normal',
use_bias=False)(se)
se = multiply([input, se])
return se
if __name__ == "__main__":
model = generate_model_2()
# train_model(model, DATASET_INDEX, dataset_prefix='kick_vs_punch_', epochs=1000, batch_size=128)
evaluate_model(model,
DATASET_INDEX,
dataset_prefix='kick_vs_punch_',
batch_size=128)
# add load model code here to fine-tune
return model
def squeeze_excite_block(input):
''' Create a squeeze-excite block
Args:
input: input tensor
filters: number of output filters
k: width factor
Returns: a keras tensor
'''
filters = input._keras_shape[-1] # channel_axis = -1 for TF
se = GlobalAveragePooling1D()(input)
se = Reshape((1, filters))(se)
se = Dense(filters // 16, activation='relu', kernel_initializer='he_normal', use_bias=False)(se)
se = Dense(filters, activation='sigmoid', kernel_initializer='he_normal', use_bias=False)(se)
se = multiply([input, se])
return se
if __name__ == "__main__":
model = generate_model_2()
# train_model(model, DATASET_INDEX, dataset_prefix='walk_vs_run_', epochs=1000, batch_size=128)
evaluate_model(model, DATASET_INDEX, dataset_prefix='walk_vs_run_', batch_size=128)
filters: number of output filters
k: width factor
Returns: a keras tensor
'''
filters = input._keras_shape[-1] # channel_axis = -1 for TF
se = GlobalAveragePooling1D()(input)
se = Reshape((1, filters))(se)
se = Dense(filters // 16,
activation='relu',
kernel_initializer='he_normal',
use_bias=False)(se)
se = Dense(filters,
activation='sigmoid',
kernel_initializer='he_normal',
use_bias=False)(se)
se = multiply([input, se])
return se
if __name__ == "__main__":
model = generate_model_2()
#train_model(model, DATASET_INDEX, dataset_prefix='action_3d', epochs=600, batch_size=128)
evaluate_model(model,
DATASET_INDEX,
dataset_prefix='action_3d',
batch_size=128)
filters: number of output filters
k: width factor
Returns: a keras tensor
'''
filters = input._keras_shape[-1] # channel_axis = -1 for TF
se = GlobalAveragePooling1D()(input)
se = Reshape((1, filters))(se)
se = Dense(filters // 16,
activation='relu',
kernel_initializer='he_normal',
use_bias=False)(se)
se = Dense(filters,
activation='sigmoid',
kernel_initializer='he_normal',
use_bias=False)(se)
se = multiply([input, se])
return se
if __name__ == "__main__":
model = generate_model_2()
# train_model(model, DATASET_INDEX, dataset_prefix='ht_sensor', epochs=600, batch_size=128)
evaluate_model(model,
DATASET_INDEX,
dataset_prefix='ht_sensor',
batch_size=128)
import json
''' Train portion '''
scores = []
for i in range(10):
K.clear_session()
print("Begin iteration %d" % (i + 1))
print("*" * 80)
print()
model = generate_model() # change to generate_model_2()
#train_model(model, DATASET_INDEX, dataset_prefix='character_attention', dataset_fold_id=(i + 1), epochs=600, batch_size=128)
score = evaluate_model(model,
DATASET_INDEX,
dataset_prefix='character_attention',
dataset_fold_id=(i + 1),
batch_size=128)
scores.append(score)
with open('data/character/scores.json', 'w') as f:
json.dump({'scores': scores}, f)
''' evaluate average score '''
with open('data/character/scores.json', 'r') as f:
results = json.load(f)
scores = results['scores']
avg_score = sum(scores) / len(scores)
print("Scores : ", scores)
print("Average score over 10 epochs : ", avg_score)
'''
filters = input._keras_shape[-1] # channel_axis = -1 for TF
se = GlobalAveragePooling1D()(input)
se = Reshape((1, filters))(se)
se = Dense(filters // 16,
activation='relu',
kernel_initializer='he_normal',
use_bias=False)(se)
se = Dense(filters,
activation='sigmoid',
kernel_initializer='he_normal',
use_bias=False)(se)
se = multiply([input, se])
return se
if __name__ == "__main__":
model = generate_model()
train_model(model,
DATASET_INDEX,
dataset_prefix='daily_sport_no_attention',
epochs=500,
batch_size=128)
evaluate_model(model,
DATASET_INDEX,
dataset_prefix='daily_sport_no_attention',
batch_size=128)
filters: number of output filters
k: width factor
Returns: a keras tensor
'''
filters = input._keras_shape[-1] # channel_axis = -1 for TF
se = GlobalAveragePooling1D()(input)
se = Reshape((1, filters))(se)
se = Dense(filters // 16,
activation='relu',
kernel_initializer='he_normal',
use_bias=False)(se)
se = Dense(filters,
activation='sigmoid',
kernel_initializer='he_normal',
use_bias=False)(se)
se = multiply([input, se])
return se
if __name__ == "__main__":
model = generate_model_2()
# train_model(model, DATASET_INDEX, dataset_prefix='net_flow_', epochs=1000, batch_size=128)
evaluate_model(model,
DATASET_INDEX,
dataset_prefix='net_flow_',
batch_size=128)
filters: number of output filters
k: width factor
Returns: a keras tensor
'''
filters = input._keras_shape[-1] # channel_axis = -1 for TF
se = GlobalAveragePooling1D()(input)
se = Reshape((1, filters))(se)
se = Dense(filters // 16,
activation='relu',
kernel_initializer='he_normal',
use_bias=False)(se)
se = Dense(filters,
activation='sigmoid',
kernel_initializer='he_normal',
use_bias=False)(se)
se = multiply([input, se])
return se
if __name__ == "__main__":
model = generate_model_2()
# train_model(model, DATASET_INDEX, dataset_prefix='cmu_subject_16_', epochs=1000, batch_size=128)
evaluate_model(model,
DATASET_INDEX,
dataset_prefix='cmu_subject_16_',
batch_size=128)
'''
filters = input._keras_shape[-1] # channel_axis = -1 for TF
se = GlobalAveragePooling1D()(input)
se = Reshape((1, filters))(se)
se = Dense(filters // 16,
activation='relu',
kernel_initializer='he_normal',
use_bias=False)(se)
se = Dense(filters,
activation='sigmoid',
kernel_initializer='he_normal',
use_bias=False)(se)
se = multiply([input, se])
return se
if __name__ == "__main__":
model = generate_model_2()
train_model(model,
DATASET_INDEX,
dataset_prefix='movement_aal',
epochs=1000,
batch_size=128)
evaluate_model(model,
DATASET_INDEX,
dataset_prefix='movement_aal',
batch_size=128)
import json
''' Train portion '''
scores = []
for i in range(10):
K.clear_session()
print("Begin iteration %d" % (i + 1))
print("*" * 80)
print()
model = generate_model() # change to generate_model_2()
#train_model(model, DATASET_INDEX, dataset_prefix='ck', dataset_fold_id=(i + 1), epochs=600, batch_size=128)
score = evaluate_model(model,
DATASET_INDEX,
dataset_prefix='ck',
dataset_fold_id=(i + 1),
batch_size=128)
scores.append(score)
with open('data/CK/scores.json', 'w') as f:
json.dump({'scores': scores}, f)
''' evaluate average score '''
with open('data/CK/scores.json', 'r') as f:
results = json.load(f)
scores = results['scores']
avg_score = sum(scores) / len(scores)
print("Scores : ", scores)
print("Average score over 10 epochs : ", avg_score)
'''
filters = input._keras_shape[-1] # channel_axis = -1 for TF
se = GlobalAveragePooling1D()(input)
se = Reshape((1, filters))(se)
se = Dense(filters // 16,
activation='relu',
kernel_initializer='he_normal',
use_bias=False)(se)
se = Dense(filters,
activation='sigmoid',
kernel_initializer='he_normal',
use_bias=False)(se)
se = multiply([input, se])
return se
if __name__ == "__main__":
model = generate_model()
train_model(model,
DATASET_INDEX,
dataset_prefix='gesture_phase',
epochs=1000,
batch_size=128)
evaluate_model(model,
DATASET_INDEX,
dataset_prefix='gesture_phase',
batch_size=128)
filters: number of output filters
k: width factor
Returns: a keras tensor
'''
filters = input._keras_shape[-1] # channel_axis = -1 for TF
se = GlobalAveragePooling1D()(input)
se = Reshape((1, filters))(se)
se = Dense(filters // 16,
activation='relu',
kernel_initializer='he_normal',
use_bias=False)(se)
se = Dense(filters,
activation='sigmoid',
kernel_initializer='he_normal',
use_bias=False)(se)
se = multiply([input, se])
return se
if __name__ == "__main__":
model = generate_model_2()
# train_model(model, DATASET_INDEX, dataset_prefix='character_trajectories_', epochs=1000, batch_size=128)
evaluate_model(model,
DATASET_INDEX,
dataset_prefix='character_trajectories_',
batch_size=128)
'''
filters = input._keras_shape[-1] # channel_axis = -1 for TF
se = GlobalAveragePooling1D()(input)
se = Reshape((1, filters))(se)
se = Dense(filters // 16,
activation='relu',
kernel_initializer='he_normal',
use_bias=False)(se)
se = Dense(filters,
activation='sigmoid',
kernel_initializer='he_normal',
use_bias=False)(se)
se = multiply([input, se])
return se
if __name__ == "__main__":
model = generate_model_2()
train_model(model,
DATASET_INDEX,
dataset_prefix='ozone',
epochs=600,
batch_size=128)
evaluate_model(model,
DATASET_INDEX,
dataset_prefix='ozone',
batch_size=128)
filters: number of output filters
k: width factor
Returns: a keras tensor
'''
filters = input._keras_shape[-1] # channel_axis = -1 for TF
se = GlobalAveragePooling1D()(input)
se = Reshape((1, filters))(se)
se = Dense(filters // 16,
activation='relu',
kernel_initializer='he_normal',
use_bias=False)(se)
se = Dense(filters,
activation='sigmoid',
kernel_initializer='he_normal',
use_bias=False)(se)
se = multiply([input, se])
return se
if __name__ == "__main__":
model = generate_model_2()
# train_model(model, DATASET_INDEX, dataset_prefix='shapes_random_', epochs=1000, batch_size=128)
evaluate_model(model,
DATASET_INDEX,
dataset_prefix='shapes_random_',
batch_size=128)
# add load model code here to fine-tune
return model
def squeeze_excite_block(input):
''' Create a squeeze-excite block
Args:
input: input tensor
filters: number of output filters
k: width factor
Returns: a keras tensor
'''
filters = input._keras_shape[-1] # channel_axis = -1 for TF
se = GlobalAveragePooling1D()(input)
se = Reshape((1, filters))(se)
se = Dense(filters // 16, activation='relu', kernel_initializer='he_normal', use_bias=False)(se)
se = Dense(filters, activation='sigmoid', kernel_initializer='he_normal', use_bias=False)(se)
se = multiply([input, se])
return se
if __name__ == "__main__":
model = generate_model_2()
# train_model(model, DATASET_INDEX, dataset_prefix='u_wave_', epochs=1000, batch_size=128)
evaluate_model(model, DATASET_INDEX, dataset_prefix='u_wave_', batch_size=128)
filters: number of output filters
k: width factor
Returns: a keras tensor
'''
filters = input._keras_shape[-1] # channel_axis = -1 for TF
se = GlobalAveragePooling1D()(input)
se = Reshape((1, filters))(se)
se = Dense(filters // 16,
activation='relu',
kernel_initializer='he_normal',
use_bias=False)(se)
se = Dense(filters,
activation='sigmoid',
kernel_initializer='he_normal',
use_bias=False)(se)
se = multiply([input, se])
return se
if __name__ == "__main__":
model = generate_model_2()
# train_model(model, DATASET_INDEX, dataset_prefix='occupancy_detect', epochs=1000, batch_size=128)
evaluate_model(model,
DATASET_INDEX,
dataset_prefix='occupancy_detect',
batch_size=128)
# add load model code here to fine-tune
return model
def squeeze_excite_block(input):
''' Create a squeeze-excite block
Args:
input: input tensor
filters: number of output filters
k: width factor
Returns: a keras tensor
'''
filters = input._keras_shape[-1] # channel_axis = -1 for TF
se = GlobalAveragePooling1D()(input)
se = Reshape((1, filters))(se)
se = Dense(filters // 16, activation='relu', kernel_initializer='he_normal', use_bias=False)(se)
se = Dense(filters, activation='sigmoid', kernel_initializer='he_normal', use_bias=False)(se)
se = multiply([input, se])
return se
if __name__ == "__main__":
model = generate_model_2()
# train_model(model, DATASET_INDEX, dataset_prefix='japanese_vowels_', epochs=1000, batch_size=128)
evaluate_model(model, DATASET_INDEX, dataset_prefix='japanese_vowels_', batch_size=128)
filters: number of output filters
k: width factor
Returns: a keras tensor
'''
filters = input._keras_shape[-1] # channel_axis = -1 for TF
se = GlobalAveragePooling1D()(input)
se = Reshape((1, filters))(se)
se = Dense(filters // 16,
activation='relu',
kernel_initializer='he_normal',
use_bias=False)(se)
se = Dense(filters,
activation='sigmoid',
kernel_initializer='he_normal',
use_bias=False)(se)
se = multiply([input, se])
return se
if __name__ == "__main__":
model = generate_model_2()
# train_model(model, DATASET_INDEX, dataset_prefix='pendigits_', epochs=1000, batch_size=128)
evaluate_model(model,
DATASET_INDEX,
dataset_prefix='pendigits_',
batch_size=128)
