def image_classifier_xception(conf, input, **kw):
# extract conf
f = conf['fit']['args']
e = conf['evaluate']['args']
epochs = f['epochs']
batch_size = f['batch_size']
# extract kw
result_sds = kw.pop('result_sds', None)
project_id = kw.pop('project_id', None)
result_dir = kw.pop('result_dir', None)
# extract input
train_data_dir = input['train_data_dir']
validation_data_dir = input['validation_data_dir']
nb_train_samples = input['nb_train_samples']
nb_validation_samples = input['nb_validation_samples']
# dimensions of our images.
# use 150, 150 as default
img_width, img_height = 150, 150
if K.image_data_format() == 'channels_first':
input_shape = (3, img_width, img_height)
else:
input_shape = (img_width, img_height, 3)
with graph.as_default():
return model_main(result_sds, project_id, result_dir, train_data_dir,
validation_data_dir, nb_train_samples,
nb_validation_samples, input_shape, img_width,
img_height, epochs, batch_size)
def keras_fmin_fnct(space):
with graph.as_default():
model = Sequential()
model.add(Dense(units=space['units'], activation=space['activation'], input_shape=[4]))
model.add(Dropout(rate=space['rate']))
model.add(Dense(units=2, activation='softmax'))
model.compile(optimizer=SGD(lr=space['lr']), loss=space['loss'], metrics=['acc'])
model.fit(x_train, y_train, validation_data=(x_test, y_test), batch_size=128, epochs=10, verbose=0)
score, acc = model.evaluate(x_test, y_test, batch_size=128)
return {'loss': -acc, 'status': STATUS_OK, 'model': model}
def mlp(conf, input, **kw):
result_sds = kw.pop('result_sds', None)
project_id = kw.pop('project_id', None)
result_dir = kw.pop('result_dir', None)
job_id = kw.pop('job_id', None)
project = project_business.get_by_id(project_id)
ow = ownership_business.get_ownership_by_owned_item(project, 'project')
user_ID = ow.user.user_ID
f = conf['fit']
e = conf['evaluate']
x_train = input['x_tr']
y_train = input['y_tr']
x_val = input['x_te']
y_val = input['y_te']
x_test = input['x_te']
y_test = input['y_te']
with graph.as_default():
return mlp_main(result_sds, project_id, job_id, user_ID, result_dir,
x_train, y_train, x_val, y_val, x_test, y_test, f, e)
def export(job_id, user_ID):
"""
export model for tf serving
:param job_id: str/ObjectId
:return:
"""
result_dir, h5_filename = get_results_dir_by_job_id(job_id, user_ID)
# result_sds = staging_data_set_business.get_by_job_id(job_id)
model_dir = os.path.join(result_dir + '/', 'model.json')
weights_dir = os.path.join(result_dir + '/', h5_filename)
with open(model_dir, 'r') as f:
data = json.load(f)
json_string = json.dumps(data)
# new_g = tf.Graph()
with graph.as_default():
model = model_from_json(json_string)
# model.load_weights(weights_dir)
# working_dir = MODEL_EXPORT_BASE
# export_base_path = os.path.join(working_dir, str(result_sds.id))
version = keras_saved_model.export(model, result_dir, weights_dir)
return result_dir, version
def mnist_irnn(conf, input, **kw):
result_sds = kw.pop('result_sds', None)
project_id = kw.pop('project_id', None)
f = conf['fit']
e = conf['evaluate']
x_train = input['x_tr']
y_train = input['y_tr']
x_val = input['x_te']
y_val = input['y_te']
x_test = input['x_te']
y_test = input['y_te']
x_train = x_train.reshape(x_train.shape[0], -1, 1)
x_test = x_test.reshape(x_test.shape[0], -1, 1)
x_val = x_test
x_train_shape = x_train.shape
input_shape = x_train_shape[1:]
num_classes = y_train.shape[1]
hidden_units = 100
learning_rate = 1e-6
with graph.as_default():
model = Sequential()
model.add(
SimpleRNN(
hidden_units,
kernel_initializer=initializers.RandomNormal(stddev=0.001),
recurrent_initializer=initializers.Identity(gain=1.0),
activation='relu',
input_shape=input_shape))
model.add(Dense(num_classes))
model.add(Activation('softmax'))
rmsprop = RMSprop(lr=learning_rate)
model.compile(loss='categorical_crossentropy',
optimizer=rmsprop,
metrics=['accuracy'])
# callback to save metrics
batch_print_callback = LambdaCallback(
on_epoch_end=lambda epoch, logs: logger_service.log_epoch_end(
epoch, logs, result_sds, project_id))
# checkpoint to save best weight
best_checkpoint = MongoModelCheckpoint(result_sds=result_sds,
verbose=0,
save_best_only=True)
# checkpoint to save latest weight
general_checkpoint = MongoModelCheckpoint(result_sds=result_sds,
verbose=0)
# training
history = model.fit(x_train,
y_train,
validation_data=(x_val, y_val),
callbacks=[
batch_print_callback, best_checkpoint,
general_checkpoint
],
verbose=0,
**f['args'])
score = model.evaluate(x_test, y_test, **e['args'])
# weights = model.get_weights()
config = model.get_config()
logger_service.log_train_end(result_sds,
model_config=config,
score=score,
history=history.history)
return {'score': score, 'history': history.history}
def mnist_mlp(conf, input, **kw):
result_sds = kw.pop('result_sds', None)
project_id = kw.pop('project_id', None)
f = conf['fit']
e = conf['evaluate']
x_train = input['x_tr']
y_train = input['y_tr']
x_val = input['x_te']
y_val = input['y_te']
x_test = input['x_te']
y_test = input['y_te']
num_classes = y_train.shape[1]
# 获取 img 的格式
x_train_shape = x_train.shape
if x_train_shape[1] > 3:
# 格式为 (None, img_rows, img_cols, 1)
x_train = x_train.reshape(-1, x_train_shape[1] * x_train_shape[2])
x_test = x_test.reshape(-1, x_train_shape[1] * x_train_shape[2])
x_val = x_test
else:
# 格式为 (None, 1, img_rows, img_cols)
x_train = x_train.reshape(-1, x_train_shape[2] * x_train_shape[3])
x_test = x_test.reshape(-1, x_train_shape[2] * x_train_shape[3])
x_val = x_test
# print(x_train.shape)
with graph.as_default():
model = Sequential()
model.add(
Dense(512, activation='relu', input_shape=(x_train.shape[1], )))
model.add(Dropout(0.2))
model.add(Dense(512, activation='relu'))
model.add(Dropout(0.2))
model.add(Dense(num_classes, activation='softmax'))
model.compile(loss='categorical_crossentropy',
optimizer=RMSprop(),
metrics=['accuracy'])
# callback to save metrics
batch_print_callback = LambdaCallback(
on_epoch_end=lambda epoch, logs: logger_service.log_epoch_end(
epoch, logs, result_sds, project_id))
# checkpoint to save best weight
best_checkpoint = MongoModelCheckpoint(result_sds=result_sds,
verbose=0,
save_best_only=True)
# checkpoint to save latest weight
general_checkpoint = MongoModelCheckpoint(result_sds=result_sds,
verbose=0)
# training
history = model.fit(x_train,
y_train,
validation_data=(x_val, y_val),
callbacks=[
batch_print_callback, best_checkpoint,
general_checkpoint
],
verbose=0,
**f['args'])
score = model.evaluate(x_test, y_test, **e['args'])
# weights = model.get_weights()
config = model.get_config()
logger_service.log_train_end(result_sds,
model_config=config,
score=score,
history=history.history)
return {'score': score, 'history': history.history}
def convnet(conf, input, **kw):
result_sds = kw.pop('result_sds', None)
project_id = kw.pop('project_id', None)
f = conf['fit']
e = conf['evaluate']
x_train = input['x_tr']
y_train = input['y_tr']
x_val = input['x_te']
y_val = input['y_te']
x_test = input['x_te']
y_test = input['y_te']
input_shape = x_train.shape[1:]
output_units = y_train.shape[-1]
with graph.as_default():
model = Sequential()
model.add(Conv2D(32, (3, 3), activation='relu', input_shape=input_shape))
model.add(Conv2D(32, (3, 3), activation='relu'))
model.add(MaxPooling2D(pool_size=(2, 2)))
model.add(Dropout(0.25))
model.add(Conv2D(64, (3, 3), activation='relu'))
model.add(Conv2D(64, (3, 3), activation='relu'))
model.add(MaxPooling2D(pool_size=(2, 2)))
model.add(Dropout(0.25))
model.add(Flatten())
model.add(Dense(256, activation='relu'))
model.add(Dropout(0.5))
model.add(Dense(output_units, activation='softmax'))
sgd = SGD(lr=0.01, decay=1e-6, momentum=0.9, nesterov=True)
model.compile(loss='categorical_crossentropy',
optimizer=sgd,
metrics=['accuracy'])
# callback to save metrics
batch_print_callback = LambdaCallback(on_epoch_end=
lambda epoch, logs:
logger_service.log_epoch_end(epoch, logs,
result_sds,
project_id))
# checkpoint to save best weight
best_checkpoint = MongoModelCheckpoint(result_sds=result_sds, verbose=0,
save_best_only=True)
# checkpoint to save latest weight
general_checkpoint = MongoModelCheckpoint(result_sds=result_sds,
verbose=0)
# training
history = model.fit(x_train, y_train,
validation_data=(x_val, y_val),
callbacks=[batch_print_callback, best_checkpoint,
general_checkpoint],
verbose=0,
**f['args'])
score = model.evaluate(x_test, y_test, **e['args'])
config = model.get_config()
logger_service.log_train_end(result_sds,
model_config=config,
score=score,
history=history.history)
return {'score': score, 'history': history.history}
def keras_seq(conf, input, **kw):
"""
a general implementation of sequential model of keras
:param conf: config dict
:return:
"""
result_sds = kw.pop('result_sds', None)
project_id = kw.pop('project_id', None)
job_id = kw.pop('job_id', None)
project = project_business.get_by_id(project_id)
ow = ownership_business.get_ownership_by_owned_item(project, 'project')
user_ID = ow.user.user_ID
print('conf')
print(conf)
result_dir = kw.pop('result_dir', None)
if result_sds is None:
raise RuntimeError('no result sds id passed to model')
if project_id is None:
raise RuntimeError('no project id passed to model')
with graph.as_default():
model = Sequential()
ls = conf['layers']
comp = conf['compile']
f = conf['fit']
e = conf['evaluate']
x_train = input['x_tr']
y_train = input['y_tr']
x_val = input['x_te']
y_val = input['y_te']
x_test = input['x_te']
y_test = input['y_te']
training_logger = logger_service.TrainingLogger(f['args']['epochs'],
project_id,
job_id,
user_ID,
result_sds)
# TODO add validator
# op = comp['optimizer']
# loop to add layers
for l in ls:
# get layer class from keras
layer_class = getattr(layers, l['name'])
# add layer
model.add(layer_class(**l['args']))
# optimiser
# sgd_class = getattr(optimizers, op['name'])
# sgd = sgd_class(**op['args'])
# define the metrics
# compile
model.compile(**comp['args'])
# callback to save metrics
batch_print_callback = LambdaCallback(on_epoch_begin=
lambda epoch, logs:
training_logger.log_epoch_begin(
epoch, logs),
on_epoch_end=
lambda epoch, logs:
training_logger.log_epoch_end(
epoch, logs),
on_batch_end=
lambda batch, logs:
training_logger.log_batch_end(
batch, logs)
)
# checkpoint to save best weight
best_checkpoint = MyModelCheckpoint(
os.path.abspath(os.path.join(result_dir, 'best.hdf5')),
save_weights_only=True,
verbose=1, save_best_only=True)
# checkpoint to save latest weight
general_checkpoint = MyModelCheckpoint(
os.path.abspath(os.path.join(result_dir, 'latest.hdf5')),
save_weights_only=True,
verbose=1)
# training
history = model.fit(x_train, y_train,
validation_data=(x_val, y_val),
callbacks=[batch_print_callback, best_checkpoint,
general_checkpoint],
verbose=0,
**f['args'])
# testing
score = model.evaluate(x_test, y_test, **e['args'])
# weights = model.get_weights()
config = model.get_config()
logger_service.log_train_end(result_sds,
model_config=config,
score=score,
history=history.history)
keras_saved_model.save_model(result_dir, model)
return {'score': score, 'history': history.history}
def neural_style_transfer(args, project_id, file_url):
# Path to the image to transform.
base_image_path = args.get('base_image_path')
# Path to the style reference image.
style_reference_image_path = args.get('style_reference_image_path')
# Prefix for the saved results.
result_prefix = args.get('result_prefix')
# Number of iterations to run.
iterations = args.get('iter', 3)
# these are the weights of the different loss components
# content_weight
content_weight = args.get('content_weight', 0.025)
# Style weight.
style_weight = args.get('style_weight', 2.0)
# Total Variation weight.
total_variation_weight = args.get('tv_weight', 1.0)
# dimensions of the generated picture.
width, height = load_img(base_image_path).size
img_nrows = 400
img_ncols = int(width * img_nrows / height)
with graph.as_default():
# this Evaluator class makes it possible
# to compute loss and gradients in one pass
# while retrieving them via two separate functions,
# "loss" and "grads". This is done because scipy.optimize
# requires separate functions for loss and gradients,
# but computing them separately would be inefficient.
class Evaluator(object):
def __init__(self):
self.loss_value = None
self.grads_values = None
def loss(self, x):
assert self.loss_value is None
loss_value, grad_values = eval_loss_and_grads(x)
self.loss_value = loss_value
self.grad_values = grad_values
return self.loss_value
def grads(self, x):
assert self.loss_value is not None
grad_values = np.copy(self.grad_values)
self.loss_value = None
self.grad_values = None
return grad_values
# util function to open, resize and format pictures into appropriate
# tensors
# 此步骤将img的channel顺序由RGB转到了BGR
# 主要是因为 vgg模型当时是用caffe训练的，使用了opencv来加载图像，
# 而opencv的加载顺序是 BGR
# 结果是 VGG 的输入图像需要转换到 BGR模式
def preprocess_image(image_path):
img = load_img(image_path, target_size=(img_nrows, img_ncols))
img = img_to_array(img)
img = np.expand_dims(img, axis=0)
img = vgg19.preprocess_input(img)
return img
# util function to convert a tensor into a valid image
# 此步骤又从BGR转换回 RGB
def deprocess_image(x):
if K.image_data_format() == 'channels_first':
x = x.reshape((3, img_nrows, img_ncols))
x = x.transpose((1, 2, 0))
else:
x = x.reshape((img_nrows, img_ncols, 3))
# Remove zero-center by mean pixel
x[:, :, 0] += 103.939
x[:, :, 1] += 116.779
x[:, :, 2] += 123.68
# 'BGR'->'RGB'
x = x[:, :, ::-1]
x = np.clip(x, 0, 255).astype('uint8')
return x
# compute the neural style loss
# first we need to define 4 util functions
# the gram matrix of an image tensor (feature-wise outer product)
def gram_matrix(x):
assert K.ndim(x) == 3
if K.image_data_format() == 'channels_first':
features = K.batch_flatten(x)
else:
features = K.batch_flatten(K.permute_dimensions(x, (2, 0, 1)))
gram = K.dot(features, K.transpose(features))
return gram
# the "style loss" is designed to maintain
# the style of the reference image in the generated image.
# It is based on the gram matrices (which capture style) of
# feature maps from the style reference image
# and from the generated image
def style_loss(style, combination):
assert K.ndim(style) == 3
assert K.ndim(combination) == 3
S = gram_matrix(style)
C = gram_matrix(combination)
channels = 3
size = img_nrows * img_ncols
return K.sum(K.square(S - C)) / (4. * (channels**2) * (size**2))
# an auxiliary loss function
# designed to maintain the "content" of the
# base image in the generated image
def content_loss(base, combination):
return K.sum(K.square(combination - base))
# the 3rd loss function, total variation loss,
# designed to keep the generated image locally coherent
def total_variation_loss(x):
assert K.ndim(x) == 4
if K.image_data_format() == 'channels_first':
a = K.square(x[:, :, :img_nrows - 1, :img_ncols - 1] -
x[:, :, 1:, :img_ncols - 1])
b = K.square(x[:, :, :img_nrows - 1, :img_ncols - 1] -
x[:, :, :img_nrows - 1, 1:])
else:
a = K.square(x[:, :img_nrows - 1, :img_ncols - 1, :] -
x[:, 1:, :img_ncols - 1, :])
b = K.square(x[:, :img_nrows - 1, :img_ncols - 1, :] -
x[:, :img_nrows - 1, 1:, :])
return K.sum(K.pow(a + b, 1.25))
def eval_loss_and_grads(x):
if K.image_data_format() == 'channels_first':
x = x.reshape((1, 3, img_nrows, img_ncols))
else:
x = x.reshape((1, img_nrows, img_ncols, 3))
outs = f_outputs([x])
loss_value = outs[0]
if len(outs[1:]) == 1:
grad_values = outs[1].flatten().astype('float64')
else:
grad_values = np.array(outs[1:]).flatten().astype('float64')
return loss_value, grad_values
# get tensor representations of our images
base_image = K.variable(preprocess_image(base_image_path))
style_reference_image = K.variable(
preprocess_image(style_reference_image_path))
# this will contain our generated image
if K.image_data_format() == 'channels_first':
combination_image = K.placeholder((1, 3, img_nrows, img_ncols))
else:
combination_image = K.placeholder((1, img_nrows, img_ncols, 3))
# combine the 3 images into a single Keras tensor
input_tensor = K.concatenate(
[base_image, style_reference_image, combination_image], axis=0)
# build the VGG16 network with our 3 images as input
# the model will be loaded with pre-trained ImageNet weights
model = vgg19.VGG19(input_tensor=input_tensor,
weights='imagenet',
include_top=False)
print('Model loaded.')
# get the symbolic outputs of each "key" layer (we gave them unique names).
outputs_dict = dict([(layer.name, layer.output)
for layer in model.layers])
# combine these loss functions into a single scalar
loss = K.variable(0.)
layer_features = outputs_dict['block5_conv2']
base_image_features = layer_features[0, :, :, :]
combination_features = layer_features[2, :, :, :]
loss += content_weight * content_loss(base_image_features,
combination_features)
feature_layers = [
'block1_conv1', 'block2_conv1', 'block3_conv1', 'block4_conv1',
'block5_conv1'
]
for layer_name in feature_layers:
layer_features = outputs_dict[layer_name]
style_reference_features = layer_features[1, :, :, :]
combination_features = layer_features[2, :, :, :]
sl = style_loss(style_reference_features, combination_features)
loss += (style_weight / len(feature_layers)) * sl
loss += total_variation_weight * total_variation_loss(
combination_image)
# get the gradients of the generated image wrt the loss
grads = K.gradients(loss, combination_image)
outputs = [loss]
if isinstance(grads, (list, tuple)):
outputs += grads
else:
outputs.append(grads)
f_outputs = K.function([combination_image], outputs)
evaluator = Evaluator()
# run scipy-based optimization (L-BFGS) over the pixels of the generated
# image
# so as to minimize the neural style loss
x = preprocess_image(base_image_path)
url = ''
for i in range(iterations):
print('Start of iteration', i)
start_time = time.time()
x, min_val, info = fmin_l_bfgs_b(evaluator.loss,
x.flatten(),
fprime=evaluator.grads,
maxfun=20)
print('Current loss value:', min_val)
# save current generated image
img = deprocess_image(x.copy())
fname = result_prefix + '_at_iteration_%d.png' % i
imsave(fname, img)
end_time = time.time()
url = file_url + 'result_at_iteration_{}.png?predict=true'.format(
i)
logger_service.emit_message_url({'url': url, 'n': i}, project_id)
print('Image saved as', fname)
print('Iteration %d completed in %ds' % (i, end_time - start_time))
return {'url': url, 'n': iterations}
def imdb_lstm(conf, input, **kw):
result_sds = kw.pop('result_sds', None)
project_id = kw.pop('project_id', None)
f = conf['fit']
e = conf['evaluate']
x_train = input['x_tr']
y_train = input['y_tr']
x_val = input['x_te']
y_val = input['y_te']
x_test = input['x_te']
y_test = input['y_te']
max_features = input['max_features']
maxlen = input['maxlen']
x_train = sequence.pad_sequences(x_train, maxlen=maxlen)
x_test = sequence.pad_sequences(x_test, maxlen=maxlen)
x_val = x_test
with graph.as_default():
model = Sequential()
model.add(Embedding(max_features, 128))
model.add(LSTM(128, dropout=0.2, recurrent_dropout=0.2))
model.add(Dense(1, activation='sigmoid'))
# try using different optimizers and different optimizer configs
model.compile(loss='binary_crossentropy',
optimizer='adam',
metrics=['accuracy'])
# callback to save metrics
batch_print_callback = LambdaCallback(
on_epoch_end=lambda epoch, logs: logger_service.log_epoch_end(
epoch, logs, result_sds, project_id))
# checkpoint to save best weight
best_checkpoint = MongoModelCheckpoint(result_sds=result_sds,
verbose=0,
save_best_only=True)
# checkpoint to save latest weight
general_checkpoint = MongoModelCheckpoint(result_sds=result_sds,
verbose=0)
# training
history = model.fit(x_train,
y_train,
validation_data=(x_val, y_val),
callbacks=[
batch_print_callback, best_checkpoint,
general_checkpoint
],
verbose=0,
**f['args'])
score = model.evaluate(x_test, y_test, **e['args'])
# weights = model.get_weights()
config = model.get_config()
logger_service.log_train_end(result_sds,
model_config=config,
score=score,
history=history.history)
return {'score': score, 'history': history.history}
def imdb_fasttext(conf, input, **kw):
result_sds = kw.pop('result_sds', None)
project_id = kw.pop('project_id', None)
f = conf['fit']
e = conf['evaluate']
x_train = input['x_tr']
y_train = input['y_tr']
x_val = input['x_te']
y_val = input['y_te']
x_test = input['x_te']
y_test = input['y_te']
# Set parameters:
# ngram_range = 2 will add bi-grams features
ngram_range = input['ngram_range']
max_features = input['max_features']
maxlen = input['maxlen']
embedding_dims = 50
def create_ngram_set(input_list, ngram_value=2):
"""
Extract a set of n-grams from a list of integers.
create_ngram_set([1, 4, 9, 4, 1, 4], ngram_value=2)
{(4, 9), (4, 1), (1, 4), (9, 4)}
create_ngram_set([1, 4, 9, 4, 1, 4], ngram_value=3)
[(1, 4, 9), (4, 9, 4), (9, 4, 1), (4, 1, 4)]
"""
return set(zip(*[input_list[i:] for i in range(ngram_value)]))
def add_ngram(sequences, token_indice, ngram_range=2):
"""
Augment the input list of list (sequences) by appending n-grams values.
Example: adding bi-gram
sequences = [[1, 3, 4, 5], [1, 3, 7, 9, 2]]
token_indice = {(1, 3): 1337, (9, 2): 42, (4, 5): 2017}
add_ngram(sequences, token_indice, ngram_range=2)
[[1, 3, 4, 5, 1337, 2017], [1, 3, 7, 9, 2, 1337, 42]]
Example: adding tri-gram
sequences = [[1, 3, 4, 5], [1, 3, 7, 9, 2]]
token_indice = {(1, 3): 1337, (9, 2): 42, (4, 5): 2017, (7, 9,
2): 2018}
add_ngram(sequences, token_indice, ngram_range=3)
[[1, 3, 4, 5, 1337], [1, 3, 7, 9, 2, 1337, 2018]]
"""
new_sequences = []
for input_list in sequences:
new_list = input_list[:]
for i in range(len(new_list) - ngram_range + 1):
for ngram_value in range(2, ngram_range + 1):
ngram = tuple(new_list[i:i + ngram_value])
if ngram in token_indice:
new_list.append(token_indice[ngram])
new_sequences.append(new_list)
return new_sequences
if ngram_range > 1:
print('Adding {}-gram features'.format(ngram_range))
# Create set of unique n-gram from the training set.
ngram_set = set()
for input_list in x_train:
for i in range(2, ngram_range + 1):
set_of_ngram = create_ngram_set(input_list, ngram_value=i)
ngram_set.update(set_of_ngram)
# Dictionary mapping n-gram token to a unique integer.
# Integer values are greater than max_features in order
# to avoid collision with existing features.
start_index = max_features + 1
token_indice = {v: k + start_index for k, v in enumerate(ngram_set)}
indice_token = {token_indice[k]: k for k in token_indice}
# max_features is the highest integer that could be found in the
# dataset.
max_features = np.max(list(indice_token.keys())) + 1
# Augmenting x_train and x_test with n-grams features
x_train = add_ngram(x_train, token_indice, ngram_range)
x_test = add_ngram(x_test, token_indice, ngram_range)
# print('Average train sequence length: {}'.format(
# np.mean(list(map(len, x_train)), dtype=int)))
# print('Average test sequence length: {}'.format(
# np.mean(list(map(len, x_test)), dtype=int)))
print('Pad sequences (samples x time)')
x_train = sequence.pad_sequences(x_train, maxlen=maxlen)
x_test = sequence.pad_sequences(x_test, maxlen=maxlen)
x_val = x_test
print('x_train shape:', x_train.shape)
print('x_test shape:', x_test.shape)
with graph.as_default():
model = Sequential()
# we start off with an efficient embedding layer which maps
# our vocab indices into embedding_dims dimensions
model.add(Embedding(max_features,
embedding_dims,
input_length=maxlen))
# we add a GlobalAveragePooling1D, which will average the embeddings
# of all words in the document
model.add(GlobalAveragePooling1D())
# We project onto a single unit output layer, and squash it with a
# sigmoid:
model.add(Dense(1, activation='sigmoid'))
model.compile(loss='binary_crossentropy',
optimizer='adam',
metrics=['accuracy'])
# callback to save metrics
batch_print_callback = LambdaCallback(on_epoch_end=
lambda epoch, logs:
logger_service.log_epoch_end(
epoch, logs,
result_sds,
project_id))
# checkpoint to save best weight
best_checkpoint = MongoModelCheckpoint(result_sds=result_sds,
verbose=0,
save_best_only=True)
# checkpoint to save latest weight
general_checkpoint = MongoModelCheckpoint(result_sds=result_sds,
verbose=0)
# training
history = model.fit(x_train, y_train,
validation_data=(x_val, y_val),
callbacks=[batch_print_callback, best_checkpoint,
general_checkpoint],
verbose=0,
**f['args'])
score = model.evaluate(x_test, y_test, **e['args'])
# weights = model.get_weights()
config = model.get_config()
logger_service.log_train_end(result_sds,
model_config=config,
score=score,
history=history.history)
return {
'score': score,
'history': history.history}
def imdb_cnn_lstm(conf, input, **kw):
result_sds = kw.pop('result_sds', None)
project_id = kw.pop('project_id', None)
f = conf['fit']
e = conf['evaluate']
x_train = input['x_tr']
y_train = input['y_tr']
x_val = input['x_te']
y_val = input['y_te']
x_test = input['x_te']
y_test = input['y_te']
# Embedding
embedding_size = 128
max_features = input['max_features']
maxlen = input['maxlen']
x_train = sequence.pad_sequences(x_train, maxlen=maxlen)
x_test = sequence.pad_sequences(x_test, maxlen=maxlen)
x_val = x_test
# set parameters:
# Convolution
kernel_size = 5
filters = 64
pool_size = 4
# LSTM
lstm_output_size = 70
with graph.as_default():
model = Sequential()
model.add(Embedding(max_features, embedding_size, input_length=maxlen))
model.add(Dropout(0.25))
model.add(Conv1D(filters,
kernel_size,
padding='valid',
activation='relu',
strides=1))
model.add(MaxPooling1D(pool_size=pool_size))
model.add(LSTM(lstm_output_size))
model.add(Dense(1))
model.add(Activation('sigmoid'))
model.compile(loss='binary_crossentropy',
optimizer='adam',
metrics=['accuracy'])
# callback to save metrics
batch_print_callback = LambdaCallback(on_epoch_end=
lambda epoch, logs:
logger_service.log_epoch_end(
epoch, logs,
result_sds,
project_id))
# checkpoint to save best weight
best_checkpoint = MongoModelCheckpoint(result_sds=result_sds,
verbose=0,
save_best_only=True)
# checkpoint to save latest weight
general_checkpoint = MongoModelCheckpoint(result_sds=result_sds,
verbose=0)
# training
history = model.fit(x_train, y_train,
validation_data=(x_val, y_val),
callbacks=[batch_print_callback, best_checkpoint,
general_checkpoint],
verbose=0,
**f['args'])
score = model.evaluate(x_test, y_test, **e['args'])
# weights = model.get_weights()
config = model.get_config()
logger_service.log_train_end(result_sds,
model_config=config,
score=score,
history=history.history)
return {
'score': score,
'history': history.history}
def imdb_cnn(conf, input, **kw):
result_sds = kw.pop('result_sds', None)
project_id = kw.pop('project_id', None)
f = conf['fit']
e = conf['evaluate']
x_train = input['x_tr']
y_train = input['y_tr']
x_val = input['x_te']
y_val = input['y_te']
x_test = input['x_te']
y_test = input['y_te']
max_features = input['max_features']
maxlen = input['maxlen']
x_train = sequence.pad_sequences(x_train, maxlen=maxlen)
x_test = sequence.pad_sequences(x_test, maxlen=maxlen)
x_val = x_test
# set parameters:
embedding_dims = 50
filters = 250
kernel_size = 3
hidden_dims = 250
with graph.as_default():
model = Sequential()
model.add(Embedding(max_features, embedding_dims, input_length=maxlen))
model.add(Dropout(0.2))
# we add a Convolution1D, which will learn filters
# word group filters of size filter_length:
model.add(
Conv1D(filters,
kernel_size,
padding='valid',
activation='relu',
strides=1))
# we use max pooling:
model.add(GlobalMaxPooling1D())
# We add a vanilla hidden layer:
model.add(Dense(hidden_dims))
model.add(Dropout(0.2))
model.add(Activation('relu'))
# We project onto a single unit output layer, and squash it with a
# sigmoid:
model.add(Dense(1))
model.add(Activation('sigmoid'))
model.compile(loss='binary_crossentropy',
optimizer='adam',
metrics=['accuracy'])
# callback to save metrics
batch_print_callback = LambdaCallback(
on_epoch_end=lambda epoch, logs: logger_service.log_epoch_end(
epoch, logs, result_sds, project_id))
# checkpoint to save best weight
best_checkpoint = MongoModelCheckpoint(result_sds=result_sds,
verbose=0,
save_best_only=True)
# checkpoint to save latest weight
general_checkpoint = MongoModelCheckpoint(result_sds=result_sds,
verbose=0)
# training
history = model.fit(x_train,
y_train,
validation_data=(x_val, y_val),
callbacks=[
batch_print_callback, best_checkpoint,
general_checkpoint
],
verbose=0,
**f['args'])
score = model.evaluate(x_test, y_test, **e['args'])
# weights = model.get_weights()
config = model.get_config()
logger_service.log_train_end(result_sds,
model_config=config,
score=score,
history=history.history)
return {'score': score, 'history': history.history}
