def remove_space(l):
labelUtil = LabelUtil.getInstance()
ret = []
for i in range(len(l)):
if l[i] != labelUtil.get_space_index():
ret.append(l[i])
return ret
def remove_space(l):
labelUtil = LabelUtil.getInstance()
ret = []
for i in range(len(l)):
if l[i] != labelUtil.get_space_index():
ret.append(l[i])
return ret
def update(self, labels, preds):
check_label_shapes(labels, preds)
log = LogUtil().getlogger()
labelUtil = LabelUtil.getInstance()
for label, pred in zip(labels, preds):
label = label.asnumpy()
pred = pred.asnumpy()
for i in range(int(int(self.batch_size) / int(self.num_gpu))):
l = remove_blank(label[i])
p = []
for k in range(int(self.seq_length)):
p.append(np.argmax(pred[k * int(int(self.batch_size) / int(self.num_gpu)) + i]))
p = pred_best(p)
l_distance = levenshtein_distance(l, p)
self.total_n_label += len(l)
self.total_l_dist += l_distance
this_cer = float(l_distance) / float(len(l))
log.info("label: %s " % (labelUtil.convert_num_to_word(l)))
log.info("pred : %s , cer: %f (distance: %d/ label length: %d)" % (
labelUtil.convert_num_to_word(p), this_cer, l_distance, len(l)))
self.num_inst += 1
self.sum_metric += this_cer
if self.is_epoch_end:
loss = ctc_loss(l, pred, i, int(self.seq_length), int(self.batch_size), int(self.num_gpu))
self.total_ctc_loss += loss
log.info("loss: %f " % loss)
def update(self, labels, preds):
check_label_shapes(labels, preds)
if self.is_logging:
log = LogUtil().getlogger()
labelUtil = LabelUtil.getInstance()
self.batch_loss = 0.
for label, pred in zip(labels, preds):
label = label.asnumpy()
pred = pred.asnumpy()
for i in range(int(int(self.batch_size) / int(self.num_gpu))):
l = remove_blank(label[i])
p = []
for k in range(int(self.seq_length)):
p.append(np.argmax(pred[k * int(int(self.batch_size) / int(self.num_gpu)) + i]))
p = pred_best(p)
l_distance = levenshtein_distance(l, p)
self.total_n_label += len(l)
self.total_l_dist += l_distance
this_cer = float(l_distance) / float(len(l))
if self.is_logging:
log.info("label: %s " % (labelUtil.convert_num_to_word(l)))
log.info("pred : %s , cer: %f (distance: %d/ label length: %d)" % (
labelUtil.convert_num_to_word(p), this_cer, l_distance, len(l)))
self.num_inst += 1
self.sum_metric += this_cer
if self.is_epoch_end:
loss = ctc_loss(l, pred, i, int(self.seq_length), int(self.batch_size), int(self.num_gpu))
self.batch_loss += loss
if self.is_logging:
log.info("loss: %f " % loss)
self.total_ctc_loss += self.batch_loss
def prepare_minibatch(self,
audio_paths,
texts,
overwrite=False,
is_bi_graphemes=False,
seq_length=-1,
save_feature_as_csvfile=False):
""" Featurize a minibatch of audio, zero pad them and return a dictionary
Params:
audio_paths (list(str)): List of paths to audio files
texts (list(str)): List of texts corresponding to the audio files
Returns:
dict: See below for contents
"""
assert len(audio_paths) == len(
texts
), "Inputs and outputs to the network must be of the same number"
# Features is a list of (timesteps, feature_dim) arrays
# Calculate the features for each audio clip, as the log of the
# Fourier Transform of the audio
features = [
self.featurize(a,
overwrite=overwrite,
save_feature_as_csvfile=save_feature_as_csvfile)
for a in audio_paths
]
input_lengths = [f.shape[0] for f in features]
feature_dim = features[0].shape[1]
mb_size = len(features)
# Pad all the inputs so that they are all the same length
if seq_length == -1:
x = np.zeros((mb_size, self.max_seq_length, feature_dim))
else:
x = np.zeros((mb_size, seq_length, feature_dim))
y = np.zeros((mb_size, self.max_label_length))
labelUtil = LabelUtil.getInstance()
label_lengths = []
for i in range(mb_size):
feat = features[i]
feat = self.normalize(feat) # Center using means and std
x[i, :feat.shape[0], :] = feat
if is_bi_graphemes:
label = generate_bi_graphemes_label(texts[i])
label = labelUtil.convert_bi_graphemes_to_num(label)
y[i, :len(label)] = label
else:
label = labelUtil.convert_word_to_num(texts[i])
y[i, :len(texts[i])] = label
label_lengths.append(len(label))
return {
'x': x, # (0-padded features of shape(mb_size,timesteps,feat_dim)
'y': y, # list(int) Flattened labels (integer sequences)
'texts': texts, # list(str) Original texts
'input_lengths': input_lengths, # list(int) Length of each input
'label_lengths': label_lengths, # list(int) Length of each label
}
def prepare_minibatch(self, audio_paths, texts, overwrite=False,
is_bi_graphemes=False, seq_length=-1, save_feature_as_csvfile=False):
""" Featurize a minibatch of audio, zero pad them and return a dictionary
Params:
audio_paths (list(str)): List of paths to audio files
texts (list(str)): List of texts corresponding to the audio files
Returns:
dict: See below for contents
"""
assert len(audio_paths) == len(texts),\
"Inputs and outputs to the network must be of the same number"
# Features is a list of (timesteps, feature_dim) arrays
# Calculate the features for each audio clip, as the log of the
# Fourier Transform of the audio
features = [self.featurize(a, overwrite=overwrite, save_feature_as_csvfile=save_feature_as_csvfile) for a in audio_paths]
input_lengths = [f.shape[0] for f in features]
feature_dim = features[0].shape[1]
mb_size = len(features)
# Pad all the inputs so that they are all the same length
if seq_length == -1:
x = np.zeros((mb_size, self.max_seq_length, feature_dim))
else:
x = np.zeros((mb_size, seq_length, feature_dim))
y = np.zeros((mb_size, self.max_label_length))
labelUtil = LabelUtil.getInstance()
label_lengths = []
for i in range(mb_size):
feat = features[i]
feat = self.normalize(feat) # Center using means and std
x[i, :feat.shape[0], :] = feat
if is_bi_graphemes:
label = generate_bi_graphemes_label(texts[i])
label = labelUtil.convert_bi_graphemes_to_num(label)
y[i, :len(label)] = label
else:
label = labelUtil.convert_word_to_num(texts[i])
y[i, :len(texts[i])] = label
label_lengths.append(len(label))
return {
'x': x, # (0-padded features of shape(mb_size,timesteps,feat_dim)
'y': y, # list(int) Flattened labels (integer sequences)
'texts': texts, # list(str) Original texts
'input_lengths': input_lengths, # list(int) Length of each input
'label_lengths': label_lengths, # list(int) Length of each label
}
mx.random.seed(hash(datetime.now()))
# set parameters from cfg file
args = parse_args(sys.argv[1])
log_filename = args.config.get('common', 'log_filename')
log = LogUtil(filename=log_filename).getlogger()
# set parameters from data section(common)
mode = args.config.get('common', 'mode')
if mode not in ['train', 'predict', 'load']:
raise Exception(
'Define mode in the cfg file first. train or predict or load can be the candidate for the mode.')
# get meta file where character to number conversions are defined
language = args.config.get('data', 'language')
labelUtil = LabelUtil.getInstance()
if language == "en":
labelUtil.load_unicode_set("resources/unicodemap_en_baidu.csv")
else:
raise Exception("Error: Language Type: %s" % language)
args.config.set('arch', 'n_classes', str(labelUtil.get_count()))
contexts = parse_contexts(args)
num_gpu = len(contexts)
batch_size = args.config.getint('common', 'batch_size')
# check the number of gpus is positive divisor of the batch size
if batch_size % num_gpu != 0:
raise Exception('num_gpu should be positive divisor of batch_size')
if mode == "predict":
def load_data(args):
mode = args.config.get('common', 'mode')
if mode not in ['train', 'predict', 'load']:
raise Exception(
'mode must be the one of the followings - train,predict,load')
batch_size = args.config.getint('common', 'batch_size')
whcs = WHCS()
whcs.width = args.config.getint('data', 'width')
whcs.height = args.config.getint('data', 'height')
whcs.channel = args.config.getint('data', 'channel')
whcs.stride = args.config.getint('data', 'stride')
save_dir = 'checkpoints'
model_name = args.config.get('common', 'prefix')
is_bi_graphemes = args.config.getboolean('common', 'is_bi_graphemes')
overwrite_meta_files = args.config.getboolean('train',
'overwrite_meta_files')
overwrite_bi_graphemes_dictionary = args.config.getboolean(
'train', 'overwrite_bi_graphemes_dictionary')
max_duration = args.config.getfloat('data', 'max_duration')
language = args.config.get('data', 'language')
log = LogUtil().getlogger()
labelUtil = LabelUtil.getInstance()
if mode == "train" or mode == "load":
data_json = args.config.get('data', 'train_json')
val_json = args.config.get('data', 'val_json')
datagen = DataGenerator(save_dir=save_dir, model_name=model_name)
datagen.load_train_data(data_json, max_duration=max_duration)
datagen.load_validation_data(val_json, max_duration=max_duration)
if is_bi_graphemes:
if not os.path.isfile(
"resources/unicodemap_en_baidu_bi_graphemes.csv"
) or overwrite_bi_graphemes_dictionary:
load_labelutil(labelUtil=labelUtil,
is_bi_graphemes=False,
language=language)
generate_bi_graphemes_dictionary(datagen.train_texts +
datagen.val_texts)
load_labelutil(labelUtil=labelUtil,
is_bi_graphemes=is_bi_graphemes,
language=language)
args.config.set('arch', 'n_classes', str(labelUtil.get_count()))
if mode == "train":
if overwrite_meta_files:
log.info("Generate mean and std from samples")
normalize_target_k = args.config.getint(
'train', 'normalize_target_k')
datagen.sample_normalize(normalize_target_k, True)
else:
log.info("Read mean and std from meta files")
datagen.get_meta_from_file(
np.loadtxt(
generate_file_path(save_dir, model_name,
'feats_mean')),
np.loadtxt(
generate_file_path(save_dir, model_name, 'feats_std')))
elif mode == "load":
# get feat_mean and feat_std to normalize dataset
datagen.get_meta_from_file(
np.loadtxt(
generate_file_path(save_dir, model_name, 'feats_mean')),
np.loadtxt(
generate_file_path(save_dir, model_name, 'feats_std')))
elif mode == 'predict':
test_json = args.config.get('data', 'test_json')
datagen = DataGenerator(save_dir=save_dir, model_name=model_name)
datagen.load_train_data(test_json, max_duration=max_duration)
labelutil = load_labelutil(labelUtil, is_bi_graphemes, language="en")
args.config.set('arch', 'n_classes', str(labelUtil.get_count()))
datagen.get_meta_from_file(
np.loadtxt(generate_file_path(save_dir, model_name, 'feats_mean')),
np.loadtxt(generate_file_path(save_dir, model_name, 'feats_std')))
is_batchnorm = args.config.getboolean('arch', 'is_batchnorm')
if batch_size == 1 and is_batchnorm and (mode == 'train'
or mode == 'load'):
raise Warning('batch size 1 is too small for is_batchnorm')
# sort file paths by its duration in ascending order to implement sortaGrad
if mode == "train" or mode == "load":
max_t_count = datagen.get_max_seq_length(partition="train")
max_label_length = \
datagen.get_max_label_length(partition="train", is_bi_graphemes=is_bi_graphemes)
elif mode == "predict":
max_t_count = datagen.get_max_seq_length(partition="test")
max_label_length = \
datagen.get_max_label_length(partition="test", is_bi_graphemes=is_bi_graphemes)
args.config.set('arch', 'max_t_count', str(max_t_count))
args.config.set('arch', 'max_label_length', str(max_label_length))
from importlib import import_module
prepare_data_template = import_module(args.config.get('arch', 'arch_file'))
init_states = prepare_data_template.prepare_data(args)
sort_by_duration = (mode == "train")
is_bucketing = args.config.getboolean('arch', 'is_bucketing')
save_feature_as_csvfile = args.config.getboolean(
'train', 'save_feature_as_csvfile')
if is_bucketing:
buckets = json.loads(args.config.get('arch', 'buckets'))
data_loaded = BucketSTTIter(
partition="train",
count=datagen.count,
datagen=datagen,
batch_size=batch_size,
num_label=max_label_length,
init_states=init_states,
seq_length=max_t_count,
width=whcs.width,
height=whcs.height,
sort_by_duration=sort_by_duration,
is_bi_graphemes=is_bi_graphemes,
buckets=buckets,
save_feature_as_csvfile=save_feature_as_csvfile)
else:
data_loaded = STTIter(partition="train",
count=datagen.count,
datagen=datagen,
batch_size=batch_size,
num_label=max_label_length,
init_states=init_states,
seq_length=max_t_count,
width=whcs.width,
height=whcs.height,
sort_by_duration=sort_by_duration,
is_bi_graphemes=is_bi_graphemes,
save_feature_as_csvfile=save_feature_as_csvfile)
if mode == 'train' or mode == 'load':
if is_bucketing:
validation_loaded = BucketSTTIter(
partition="validation",
count=datagen.val_count,
datagen=datagen,
batch_size=batch_size,
num_label=max_label_length,
init_states=init_states,
seq_length=max_t_count,
width=whcs.width,
height=whcs.height,
sort_by_duration=False,
is_bi_graphemes=is_bi_graphemes,
buckets=buckets,
save_feature_as_csvfile=save_feature_as_csvfile)
else:
validation_loaded = STTIter(
partition="validation",
count=datagen.val_count,
datagen=datagen,
batch_size=batch_size,
num_label=max_label_length,
init_states=init_states,
seq_length=max_t_count,
width=whcs.width,
height=whcs.height,
sort_by_duration=False,
is_bi_graphemes=is_bi_graphemes,
save_feature_as_csvfile=save_feature_as_csvfile)
return data_loaded, validation_loaded, args
elif mode == 'predict':
return data_loaded, args
def load_data(args):
mode = args.config.get('common', 'mode')
batch_size = args.config.getint('common', 'batch_size')
whcs = WHCS()
whcs.width = args.config.getint('data', 'width')
whcs.height = args.config.getint('data', 'height')
whcs.channel = args.config.getint('data', 'channel')
whcs.stride = args.config.getint('data', 'stride')
save_dir = 'checkpoints'
model_name = args.config.get('common', 'prefix')
is_bi_graphemes = args.config.getboolean('common', 'is_bi_graphemes')
overwrite_meta_files = args.config.getboolean('train',
'overwrite_meta_files')
if mode == 'predict':
test_json = args.config.get('data', 'test_json')
datagen = DataGenerator(save_dir=save_dir, model_name=model_name)
datagen.load_train_data(test_json)
datagen.get_meta_from_file(
np.loadtxt(generate_file_path(save_dir, model_name, 'feats_mean')),
np.loadtxt(generate_file_path(save_dir, model_name, 'feats_std')))
elif mode == "train" or mode == "load":
data_json = args.config.get('data', 'train_json')
val_json = args.config.get('data', 'val_json')
datagen = DataGenerator(save_dir=save_dir, model_name=model_name)
datagen.load_train_data(data_json)
#test bigramphems
language = args.config.get('data', 'language')
is_bi_graphemes = args.config.getboolean('common', 'is_bi_graphemes')
if overwrite_meta_files and is_bi_graphemes:
generate_bi_graphemes_dictionary(datagen.train_texts)
labelUtil = LabelUtil.getInstance()
if language == "en":
if is_bi_graphemes:
try:
labelUtil.load_unicode_set(
"resources/unicodemap_en_baidu_bi_graphemes.csv")
except:
raise Exception(
"There is no resources/unicodemap_en_baidu_bi_graphemes.csv. Please set overwrite_meta_files at train section True"
)
else:
labelUtil.load_unicode_set("resources/unicodemap_en_baidu.csv")
else:
raise Exception("Error: Language Type: %s" % language)
args.config.set('arch', 'n_classes', str(labelUtil.get_count()))
if mode == "train":
if overwrite_meta_files:
normalize_target_k = args.config.getint(
'train', 'normalize_target_k')
datagen.sample_normalize(normalize_target_k, True)
else:
datagen.get_meta_from_file(
np.loadtxt(
generate_file_path(save_dir, model_name,
'feats_mean')),
np.loadtxt(
generate_file_path(save_dir, model_name, 'feats_std')))
datagen.load_validation_data(val_json)
elif mode == "load":
# get feat_mean and feat_std to normalize dataset
datagen.get_meta_from_file(
np.loadtxt(
generate_file_path(save_dir, model_name, 'feats_mean')),
np.loadtxt(
generate_file_path(save_dir, model_name, 'feats_std')))
datagen.load_validation_data(val_json)
else:
raise Exception(
'Define mode in the cfg file first. train or predict or load can be the candidate for the mode.'
)
is_batchnorm = args.config.getboolean('arch', 'is_batchnorm')
if batch_size == 1 and is_batchnorm:
raise Warning('batch size 1 is too small for is_batchnorm')
# sort file paths by its duration in ascending order to implement sortaGrad
if mode == "train" or mode == "load":
max_t_count = datagen.get_max_seq_length(partition="train")
max_label_length = datagen.get_max_label_length(
partition="train", is_bi_graphemes=is_bi_graphemes)
elif mode == "predict":
max_t_count = datagen.get_max_seq_length(partition="test")
max_label_length = datagen.get_max_label_length(
partition="test", is_bi_graphemes=is_bi_graphemes)
else:
raise Exception(
'Define mode in the cfg file first. train or predict or load can be the candidate for the mode.'
)
args.config.set('arch', 'max_t_count', str(max_t_count))
args.config.set('arch', 'max_label_length', str(max_label_length))
from importlib import import_module
prepare_data_template = import_module(args.config.get('arch', 'arch_file'))
init_states = prepare_data_template.prepare_data(args)
if mode == "train":
sort_by_duration = True
else:
sort_by_duration = False
data_loaded = STTIter(partition="train",
count=datagen.count,
datagen=datagen,
batch_size=batch_size,
num_label=max_label_length,
init_states=init_states,
seq_length=max_t_count,
width=whcs.width,
height=whcs.height,
sort_by_duration=sort_by_duration,
is_bi_graphemes=is_bi_graphemes)
if mode == 'predict':
return data_loaded, args
else:
validation_loaded = STTIter(partition="validation",
count=datagen.val_count,
datagen=datagen,
batch_size=batch_size,
num_label=max_label_length,
init_states=init_states,
seq_length=max_t_count,
width=whcs.width,
height=whcs.height,
sort_by_duration=False,
is_bi_graphemes=is_bi_graphemes)
return data_loaded, validation_loaded, args
def load_data(args):
mode = args.config.get('common', 'mode')
batch_size = args.config.getint('common', 'batch_size')
whcs = WHCS()
whcs.width = args.config.getint('data', 'width')
whcs.height = args.config.getint('data', 'height')
whcs.channel = args.config.getint('data', 'channel')
whcs.stride = args.config.getint('data', 'stride')
save_dir = 'checkpoints'
model_name = args.config.get('common', 'prefix')
is_bi_graphemes = args.config.getboolean('common', 'is_bi_graphemes')
overwrite_meta_files = args.config.getboolean('train', 'overwrite_meta_files')
language = args.config.get('data', 'language')
is_bi_graphemes = args.config.getboolean('common', 'is_bi_graphemes')
labelUtil = LabelUtil.getInstance()
if language == "en":
if is_bi_graphemes:
try:
labelUtil.load_unicode_set("resources/unicodemap_en_baidu_bi_graphemes.csv")
except:
raise Exception("There is no resources/unicodemap_en_baidu_bi_graphemes.csv. Please set overwrite_meta_files at train section True")
else:
labelUtil.load_unicode_set("resources/unicodemap_en_baidu.csv")
else:
raise Exception("Error: Language Type: %s" % language)
args.config.set('arch', 'n_classes', str(labelUtil.get_count()))
if mode == 'predict':
test_json = args.config.get('data', 'test_json')
datagen = DataGenerator(save_dir=save_dir, model_name=model_name)
datagen.load_train_data(test_json)
datagen.get_meta_from_file(np.loadtxt(generate_file_path(save_dir, model_name, 'feats_mean')),
np.loadtxt(generate_file_path(save_dir, model_name, 'feats_std')))
elif mode =="train" or mode == "load":
data_json = args.config.get('data', 'train_json')
val_json = args.config.get('data', 'val_json')
datagen = DataGenerator(save_dir=save_dir, model_name=model_name)
datagen.load_train_data(data_json)
#test bigramphems
if overwrite_meta_files and is_bi_graphemes:
generate_bi_graphemes_dictionary(datagen.train_texts)
args.config.set('arch', 'n_classes', str(labelUtil.get_count()))
if mode == "train":
if overwrite_meta_files:
normalize_target_k = args.config.getint('train', 'normalize_target_k')
datagen.sample_normalize(normalize_target_k, True)
else:
datagen.get_meta_from_file(np.loadtxt(generate_file_path(save_dir, model_name, 'feats_mean')),
np.loadtxt(generate_file_path(save_dir, model_name, 'feats_std')))
datagen.load_validation_data(val_json)
elif mode == "load":
# get feat_mean and feat_std to normalize dataset
datagen.get_meta_from_file(np.loadtxt(generate_file_path(save_dir, model_name, 'feats_mean')),
np.loadtxt(generate_file_path(save_dir, model_name, 'feats_std')))
datagen.load_validation_data(val_json)
else:
raise Exception(
'Define mode in the cfg file first. train or predict or load can be the candidate for the mode.')
is_batchnorm = args.config.getboolean('arch', 'is_batchnorm')
if batch_size == 1 and is_batchnorm:
raise Warning('batch size 1 is too small for is_batchnorm')
# sort file paths by its duration in ascending order to implement sortaGrad
if mode == "train" or mode == "load":
max_t_count = datagen.get_max_seq_length(partition="train")
max_label_length = datagen.get_max_label_length(partition="train",is_bi_graphemes=is_bi_graphemes)
elif mode == "predict":
max_t_count = datagen.get_max_seq_length(partition="test")
max_label_length = datagen.get_max_label_length(partition="test",is_bi_graphemes=is_bi_graphemes)
else:
raise Exception(
'Define mode in the cfg file first. train or predict or load can be the candidate for the mode.')
args.config.set('arch', 'max_t_count', str(max_t_count))
args.config.set('arch', 'max_label_length', str(max_label_length))
from importlib import import_module
prepare_data_template = import_module(args.config.get('arch', 'arch_file'))
init_states = prepare_data_template.prepare_data(args)
if mode == "train":
sort_by_duration = True
else:
sort_by_duration = False
data_loaded = STTIter(partition="train",
count=datagen.count,
datagen=datagen,
batch_size=batch_size,
num_label=max_label_length,
init_states=init_states,
seq_length=max_t_count,
width=whcs.width,
height=whcs.height,
sort_by_duration=sort_by_duration,
is_bi_graphemes=is_bi_graphemes)
if mode == 'predict':
return data_loaded, args
else:
validation_loaded = STTIter(partition="validation",
count=datagen.val_count,
datagen=datagen,
batch_size=batch_size,
num_label=max_label_length,
init_states=init_states,
seq_length=max_t_count,
width=whcs.width,
height=whcs.height,
sort_by_duration=False,
is_bi_graphemes=is_bi_graphemes)
return data_loaded, validation_loaded, args
def load_data(args):
mode = args.config.get('common', 'mode')
if mode not in ['train', 'predict', 'load']:
raise Exception('mode must be the one of the followings - train,predict,load')
batch_size = args.config.getint('common', 'batch_size')
whcs = WHCS()
whcs.width = args.config.getint('data', 'width')
whcs.height = args.config.getint('data', 'height')
whcs.channel = args.config.getint('data', 'channel')
whcs.stride = args.config.getint('data', 'stride')
save_dir = 'checkpoints'
model_name = args.config.get('common', 'prefix')
is_bi_graphemes = args.config.getboolean('common', 'is_bi_graphemes')
overwrite_meta_files = args.config.getboolean('train', 'overwrite_meta_files')
overwrite_bi_graphemes_dictionary = args.config.getboolean('train', 'overwrite_bi_graphemes_dictionary')
max_duration = args.config.getfloat('data', 'max_duration')
language = args.config.get('data', 'language')
log = LogUtil().getlogger()
labelUtil = LabelUtil.getInstance()
if mode == "train" or mode == "load":
data_json = args.config.get('data', 'train_json')
val_json = args.config.get('data', 'val_json')
datagen = DataGenerator(save_dir=save_dir, model_name=model_name)
datagen.load_train_data(data_json, max_duration=max_duration)
datagen.load_validation_data(val_json, max_duration=max_duration)
if is_bi_graphemes:
if not os.path.isfile("resources/unicodemap_en_baidu_bi_graphemes.csv") or overwrite_bi_graphemes_dictionary:
load_labelutil(labelUtil=labelUtil, is_bi_graphemes=False, language=language)
generate_bi_graphemes_dictionary(datagen.train_texts+datagen.val_texts)
load_labelutil(labelUtil=labelUtil, is_bi_graphemes=is_bi_graphemes, language=language)
args.config.set('arch', 'n_classes', str(labelUtil.get_count()))
if mode == "train":
if overwrite_meta_files:
log.info("Generate mean and std from samples")
normalize_target_k = args.config.getint('train', 'normalize_target_k')
datagen.sample_normalize(normalize_target_k, True)
else:
log.info("Read mean and std from meta files")
datagen.get_meta_from_file(
np.loadtxt(generate_file_path(save_dir, model_name, 'feats_mean')),
np.loadtxt(generate_file_path(save_dir, model_name, 'feats_std')))
elif mode == "load":
# get feat_mean and feat_std to normalize dataset
datagen.get_meta_from_file(
np.loadtxt(generate_file_path(save_dir, model_name, 'feats_mean')),
np.loadtxt(generate_file_path(save_dir, model_name, 'feats_std')))
elif mode == 'predict':
test_json = args.config.get('data', 'test_json')
datagen = DataGenerator(save_dir=save_dir, model_name=model_name)
datagen.load_train_data(test_json, max_duration=max_duration)
labelutil = load_labelutil(labelUtil, is_bi_graphemes, language="en")
args.config.set('arch', 'n_classes', str(labelUtil.get_count()))
datagen.get_meta_from_file(
np.loadtxt(generate_file_path(save_dir, model_name, 'feats_mean')),
np.loadtxt(generate_file_path(save_dir, model_name, 'feats_std')))
is_batchnorm = args.config.getboolean('arch', 'is_batchnorm')
if batch_size == 1 and is_batchnorm and (mode == 'train' or mode == 'load'):
raise Warning('batch size 1 is too small for is_batchnorm')
# sort file paths by its duration in ascending order to implement sortaGrad
if mode == "train" or mode == "load":
max_t_count = datagen.get_max_seq_length(partition="train")
max_label_length = \
datagen.get_max_label_length(partition="train", is_bi_graphemes=is_bi_graphemes)
elif mode == "predict":
max_t_count = datagen.get_max_seq_length(partition="test")
max_label_length = \
datagen.get_max_label_length(partition="test", is_bi_graphemes=is_bi_graphemes)
args.config.set('arch', 'max_t_count', str(max_t_count))
args.config.set('arch', 'max_label_length', str(max_label_length))
from importlib import import_module
prepare_data_template = import_module(args.config.get('arch', 'arch_file'))
init_states = prepare_data_template.prepare_data(args)
sort_by_duration = (mode == "train")
is_bucketing = args.config.getboolean('arch', 'is_bucketing')
save_feature_as_csvfile = args.config.getboolean('train', 'save_feature_as_csvfile')
if is_bucketing:
buckets = json.loads(args.config.get('arch', 'buckets'))
data_loaded = BucketSTTIter(partition="train",
count=datagen.count,
datagen=datagen,
batch_size=batch_size,
num_label=max_label_length,
init_states=init_states,
seq_length=max_t_count,
width=whcs.width,
height=whcs.height,
sort_by_duration=sort_by_duration,
is_bi_graphemes=is_bi_graphemes,
buckets=buckets,
save_feature_as_csvfile=save_feature_as_csvfile)
else:
data_loaded = STTIter(partition="train",
count=datagen.count,
datagen=datagen,
batch_size=batch_size,
num_label=max_label_length,
init_states=init_states,
seq_length=max_t_count,
width=whcs.width,
height=whcs.height,
sort_by_duration=sort_by_duration,
is_bi_graphemes=is_bi_graphemes,
save_feature_as_csvfile=save_feature_as_csvfile)
if mode == 'train' or mode == 'load':
if is_bucketing:
validation_loaded = BucketSTTIter(partition="validation",
count=datagen.val_count,
datagen=datagen,
batch_size=batch_size,
num_label=max_label_length,
init_states=init_states,
seq_length=max_t_count,
width=whcs.width,
height=whcs.height,
sort_by_duration=False,
is_bi_graphemes=is_bi_graphemes,
buckets=buckets,
save_feature_as_csvfile=save_feature_as_csvfile)
else:
validation_loaded = STTIter(partition="validation",
count=datagen.val_count,
datagen=datagen,
batch_size=batch_size,
num_label=max_label_length,
init_states=init_states,
seq_length=max_t_count,
width=whcs.width,
height=whcs.height,
sort_by_duration=False,
is_bi_graphemes=is_bi_graphemes,
save_feature_as_csvfile=save_feature_as_csvfile)
return data_loaded, validation_loaded, args
elif mode == 'predict':
return data_loaded, args