def generate_dataset(data_dir):
"""Generate a speech dataset."""
audio_conf = dataset.AudioConfig(sample_rate=flags_obj.sample_rate,
window_ms=flags_obj.window_ms,
stride_ms=flags_obj.stride_ms,
normalize=True)
train_data_conf = dataset.DatasetConfig(audio_conf, data_dir,
flags_obj.vocabulary_file,
flags_obj.sortagrad)
speech_dataset = dataset.DeepSpeechDataset(train_data_conf)
return speech_dataset
def generate_dataset(data_dir):
"""Generate a speech dataset."""
audio_conf = dataset.AudioConfig(flags_obj.sample_rate,
flags_obj.frame_length,
flags_obj.frame_step)
train_data_conf = dataset.DatasetConfig(
audio_conf,
data_dir,
flags_obj.vocabulary_file,
)
speech_dataset = dataset.DeepSpeechDataset(train_data_conf)
return speech_dataset
def generate_dataset(data_dir):
"""Generate a speech dataset."""
audio_conf = dataset.AudioConfig(
sample_rate=16000,
window_ms=20,
stride_ms=10,
normalize=True,
)
train_data_conf = dataset.DatasetConfig(
audio_conf, data_dir,
"C:/Users/MariamDesouky/Desktop/tuda-tpu/data/vocabulary.txt", True)
speech_dataset = dataset.DeepSpeechDataset(train_data_conf)
return speech_dataset
