tensorflow因为版本，需要修改一些方法问题

参考博文：

http://blog.csdn.net/lijjianqing/article/details/70155905

http://blog.csdn.net/caimouse/article/details/61208940

http://blog.csdn.net/edwards_june/article/details/65652385

1、AttributeError: 'module' object has no attribute 'histogram_summary'

把方法改为：  tf.summaries.histogram()，传入的参数不变

2、ValueError: Only call `softmax_cross_entropy_with_logits` with named arguments (labels=..., logits=..., ...)

xent = tf.nn.softmax_cross_entropy_with_logits(logits,
                                               labels,
                                               name="xent_raw")
改为
xent = tf.nn.softmax_cross_entropy_with_logits(logits=logits,
                                               labels=labels,
                                               name="xent_raw")

3、AttributeError: module 'tensorflow' has no attribute 'scalar_summary'

tf.image_summary('images', images)
改为
tf.summary.image('images', images)

4、AttributeError: module 'tensorflow' has no attribute 'merge_all_summaries'
self._summaries = tf.merge_all_summaries() 
改为
self._summaries = tf.summary.merge_all()