def clone_model(model, custom_objects={}):
config = {
'class_name': model.__class__.__name__,
'config': model.get_config(),
}
clone = model_from_config(config, custom_objects=custom_objects)
clone.set_weights(model.get_weights())
return clone
def clone_model(model, custom_objects={}):
# Requires Keras 1.0.7 since get_config has breaking changes.
config = {
'class_name': model.__class__.__name__,
'config': model.get_config(),
}
clone = model_from_config(config, custom_objects=custom_objects)
clone.set_weights(model.get_weights())
return clone
def load_model(path, custom_objects={}, verbose=0):
from tensorflow.keras.models import model_from_config
import json
path = splitext(path)[0]
with open('%s.json' % path, 'r') as json_file:
model_json = json_file.read()
config = json.loads(model_json)
model = model_from_config(config, custom_objects=custom_objects)
model.load_weights('%s.h5' % path)
if verbose: print('Loaded from %s' % path)
return model
def layer_from_config(config):
if 'class_name' not in config:
raise ValueError('class_name is needed to define layer')
if 'config' not in config:
# auto add empty config for layer with only class_name
config['config'] = {}
if 'name' not in config['config'] and 'name' in config:
config['config']['name'] = config['name']
return model_from_config(config,
custom_objects={
**Layers().layers,
**Activations().activations
})
def __init__(self, model, instance_state, tuner_state, metrics_config,
cloudservice):
self.instance_state = instance_state
self.tuner_state = tuner_state
self.cloudservice = cloudservice
self.state = ExecutionState(tuner_state.max_epochs, metrics_config)
# Model rereation
config = serialize_keras_object(model)
self.model = model_from_config(config)
optimizer_config = tf.keras.optimizers.serialize(model.optimizer)
optimizer = tf.keras.optimizers.deserialize(optimizer_config)
self.model.compile(optimizer=optimizer,
metrics=model.metrics,
loss=model.loss,
loss_weights=model.loss_weights)