本文整理汇总了Python中monitor.Monitor.save方法的典型用法代码示例。如果您正苦于以下问题：Python Monitor.save方法的具体用法？Python Monitor.save怎么用？Python Monitor.save使用的例子？那么恭喜您, 这里精选的方法代码示例或许可以为您提供帮助。您也可以进一步了解该方法所在类monitor.Monitor的用法示例。
在下文中一共展示了Monitor.save方法的1个代码示例，这些例子默认根据受欢迎程度排序。您可以为喜欢或者感觉有用的代码点赞，您的评价将有助于系统推荐出更棒的Python代码示例。
示例1: str
# 需要导入模块: from monitor import Monitor [as 别名]
# 或者: from monitor.Monitor import save [as 别名]
mtr.load()
# then start train with function
while True:
epoch_no = 0
while epoch_no < 100:
epoch_no += 1
print "training {} epoch, total data size is {}".format(epoch_no, str(len(data_generator.data)))
data_generator.shuffle()
cost_set = []
index = 0
for datapoint in data_generator.get_data_stream():
index += 1
train_data_sentence = datapoint[0]
train_data_result = datapoint[1]
train_data_image_feature = datapoint[2]
output, cost = train_func(train_data_sentence, train_data_result, train_data_image_feature)
cost_set.append(cost)
sys.stderr.write("train {0} data, cost = {1} \r".format(index, cost))
sys.stderr.flush()
if args.show:
print cost
print data_generator.translate(output)
print data_generator.translate(train_data_sentence)
batch_cost = sum(cost_set)/len(cost_set)
print 'current batch cost = {}'.format(batch_cost)
# save after trained
mtr.save(batch_cost)
print 'min batch cost = {}'.format(mtr.min_cost)