def __init__(self):
#### Construct the model ####
observation = cntk.ops.input_variable(STATE_DIM, np.float32, name="s")
q_target = cntk.ops.input_variable(NUM_ACTIONS, np.float32, name="q")
# Define the structure of the neural network
self.model = self.create_multi_layer_neural_network(
observation, NUM_ACTIONS, 2)
#### Define the trainer ####
self.learning_rate = cntk.learner.training_parameter_schedule(
0.0001, cntk.UnitType.sample)
self.momentum = cntk.learner.momentum_as_time_constant_schedule(0.99)
self.loss = cntk.ops.reduce_mean(cntk.ops.square(self.model -
q_target),
axis=0)
mean_error = cntk.ops.reduce_mean(cntk.ops.square(self.model -
q_target),
axis=0)
learner = cntk.adam_sgd(self.model.parameters,
self.learning_rate,
momentum=self.momentum)
self.trainer = cntk.Trainer(self.model, self.loss, mean_error, learner)
def __call__(self, parameters, opt_learning_rate=0.001, **kwargs):
lr_per_minibatch = cntk.learning_rate_schedule(
lr=opt_learning_rate, unit=cntk.UnitType.minibatch)
momentum = cntk.momentum_schedule(momentum=0.99)
return cntk.adam_sgd(parameters=parameters,
lr=lr_per_minibatch,
momentum=momentum)
def test_htk_deserializers():
mbsize = 640
epoch_size = 1000 * mbsize
lr = [0.001]
feature_dim = 33
num_classes = 132
context = 2
os.chdir(data_path)
features_file = "glob_0000.scp"
labels_file = "glob_0000.mlf"
label_mapping_file = "state.list"
fd = HTKFeatureDeserializer(
StreamDefs(amazing_features=StreamDef(
shape=feature_dim, context=(context, context), scp=features_file)))
ld = HTKMLFDeserializer(
label_mapping_file,
StreamDefs(
awesome_labels=StreamDef(shape=num_classes, mlf=labels_file)))
reader = MinibatchSource([fd, ld])
features = C.input_variable(((2 * context + 1) * feature_dim))
labels = C.input_variable((num_classes))
model = Sequential(
[For(range(3), lambda: Recurrence(LSTM(256))),
Dense(num_classes)])
z = model(features)
ce = C.cross_entropy_with_softmax(z, labels)
errs = C.classification_error(z, labels)
learner = C.adam_sgd(z.parameters,
lr=C.learning_rate_schedule(lr, C.UnitType.sample,
epoch_size),
momentum=C.momentum_as_time_constant_schedule(1000),
low_memory=True,
gradient_clipping_threshold_per_sample=15,
gradient_clipping_with_truncation=True)
trainer = C.Trainer(z, (ce, errs), learner)
input_map = {
features: reader.streams.amazing_features,
labels: reader.streams.awesome_labels
}
pp = C.ProgressPrinter(freq=0)
# just run and verify it doesn't crash
for i in range(3):
mb_data = reader.next_minibatch(mbsize, input_map=input_map)
trainer.train_minibatch(mb_data)
pp.update_with_trainer(trainer, with_metric=True)
assert True
os.chdir(abs_path)
def test_htk_deserializers():
mbsize = 640
epoch_size = 1000 * mbsize
lr = [0.001]
feature_dim = 33
num_classes = 132
context = 2
os.chdir(data_path)
features_file = "glob_0000.scp"
labels_file = "glob_0000.mlf"
label_mapping_file = "state.list"
fd = HTKFeatureDeserializer(StreamDefs(
amazing_features = StreamDef(shape=feature_dim, context=(context,context), scp=features_file)))
ld = HTKMLFDeserializer(label_mapping_file, StreamDefs(
awesome_labels = StreamDef(shape=num_classes, mlf=labels_file)))
reader = MinibatchSource([fd,ld])
features = C.input_variable(((2*context+1)*feature_dim))
labels = C.input_variable((num_classes))
model = Sequential([For(range(3), lambda : Recurrence(LSTM(256))),
Dense(num_classes)])
z = model(features)
ce = C.cross_entropy_with_softmax(z, labels)
errs = C.classification_error (z, labels)
learner = C.adam_sgd(z.parameters,
lr=C.learning_rate_schedule(lr, C.UnitType.sample, epoch_size),
momentum=C.momentum_as_time_constant_schedule(1000),
low_memory=True,
gradient_clipping_threshold_per_sample=15, gradient_clipping_with_truncation=True)
trainer = C.Trainer(z, (ce, errs), learner)
input_map={ features: reader.streams.amazing_features, labels: reader.streams.awesome_labels }
pp = C.ProgressPrinter(freq=0)
# just run and verify it doesn't crash
for i in range(3):
mb_data = reader.next_minibatch(mbsize, input_map=input_map)
trainer.train_minibatch(mb_data)
pp.update_with_trainer(trainer, with_metric=True)
assert True
os.chdir(abs_path)
def __init__(self):
#### Construct the model ####
observation = cntk.ops.input_variable(STATE_DIM, np.float32, name="s")
q_target = cntk.ops.input_variable(NUM_ACTIONS, np.float32, name="q")
# Define the structure of the neural network
self.model = self.create_multi_layer_neural_network(observation, NUM_ACTIONS, 3)
#### Define the trainer ####
self.learning_rate = 0.00025
self.loss = cntk.ops.reduce_mean(cntk.ops.square(self.model - q_target), axis=0)
mean_error = cntk.ops.reduce_mean(cntk.ops.square(self.model - q_target), axis=0)
learner = cntk.adam_sgd(self.model.parameters, self.learning_rate/self.BATCH_SIZE, momentum=0.9)
self.trainer = cntk.Trainer(self.model, self.loss, mean_error, learner)
# Define the label as the other input parameter of the trainer
labels = cntk.ops.input_variable(output_dim, np.float32)
# Initialize the parameters for the trainer
train_minibatch_size = 50
learning_rate = cntk.learner.training_parameter_schedule(1e-4, cntk.UnitType.sample)
momentum = cntk.learner.momentum_as_time_constant_schedule(0.99)
# Define the loss function
loss = cntk.ops.cross_entropy_with_softmax(output, labels)
# Define the function that calculates classification error
label_error = cntk.ops.classification_error(output, labels)
# Instantiate the trainer object to drive the model training
learner = cntk.adam_sgd(output.parameters, learning_rate, momentum)
trainer = cntk.Trainer(output, loss, label_error, [learner])
'''
-----------------------------------------
Training the Convolutional Neural Network
-----------------------------------------
'''
num_training_epoch = 20
training_progress_output_freq = 100
training_start_time = perf_counter()
for epoch in range(num_training_epoch):
sample_count = 0