2018-05-11 18:42:58 +08:00
|
|
|
|
#!/usr/bin/env python3
|
|
|
|
|
|
# -*- coding: utf-8 -*-
|
|
|
|
|
|
"""
|
|
|
|
|
|
@author: nl8590687
|
|
|
|
|
|
用于训练语音识别系统语音模型的程序
|
|
|
|
|
|
|
|
|
|
|
|
"""
|
|
|
|
|
|
import platform as plat
|
|
|
|
|
|
import os
|
|
|
|
|
|
|
|
|
|
|
|
import tensorflow as tf
|
|
|
|
|
|
from keras.backend.tensorflow_backend import set_session
|
|
|
|
|
|
|
|
|
|
|
|
|
2018-07-27 14:31:48 +08:00
|
|
|
|
from SpeechModel251 import ModelSpeech
|
2018-05-11 18:42:58 +08:00
|
|
|
|
|
|
|
|
|
|
os.environ["CUDA_VISIBLE_DEVICES"] = "0"
|
2018-07-15 14:21:10 +08:00
|
|
|
|
#进行配置，使用95%的GPU
|
2018-05-11 18:42:58 +08:00
|
|
|
|
config = tf.ConfigProto()
|
2018-07-15 14:21:10 +08:00
|
|
|
|
config.gpu_options.per_process_gpu_memory_fraction = 0.95
|
2018-05-11 18:42:58 +08:00
|
|
|
|
#config.gpu_options.allow_growth=True #不全部占满显存, 按需分配
|
|
|
|
|
|
set_session(tf.Session(config=config))
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
datapath = ''
|
|
|
|
|
|
modelpath = 'model_speech'
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
if(not os.path.exists(modelpath)): # 判断保存模型的目录是否存在
|
|
|
|
|
|
os.makedirs(modelpath) # 如果不存在，就新建一个，避免之后保存模型的时候炸掉
|
|
|
|
|
|
|
|
|
|
|
|
system_type = plat.system() # 由于不同的系统的文件路径表示不一样，需要进行判断
|
|
|
|
|
|
if(system_type == 'Windows'):
|
|
|
|
|
|
datapath = 'E:\\语音数据集'
|
|
|
|
|
|
modelpath = modelpath + '\\'
|
|
|
|
|
|
elif(system_type == 'Linux'):
|
|
|
|
|
|
datapath = 'dataset'
|
|
|
|
|
|
modelpath = modelpath + '/'
|
|
|
|
|
|
else:
|
|
|
|
|
|
print('*[Message] Unknown System\n')
|
|
|
|
|
|
datapath = 'dataset'
|
|
|
|
|
|
modelpath = modelpath + '/'
|
|
|
|
|
|
|
|
|
|
|
|
ms = ModelSpeech(datapath)
|
|
|
|
|
|
|
2018-07-27 14:31:48 +08:00
|
|
|
|
#ms.LoadModel(modelpath + 'speech_model251_e_0_step_327500.model')
|
2018-07-15 14:21:10 +08:00
|
|
|
|
ms.TrainModel(datapath, epoch = 50, batch_size = 16, save_step = 500)
|
2018-05-11 18:42:58 +08:00
|
|
|
|
|
|
|
|
|
|
|
