tensorflow版本升级报错+解决方案汇总

---

最近基于之前的大佬写过的代码，用tensorflow在写一个聊天机器人，因为版本的原因，收集到了很多报错信息，将解决方法汇总了一下，方便以后查看。

• AttributeError: ‘module’ object hasno attribute ‘histogram_summary’

  解决方法：tf.histogram_summary() 改为： tf.summaries.histogram()

• ValueError: Only call sigmoid_cross_entropy_with_logits with named arguments (labels=…, logits=…, …)

  解决方法：

  loss = tf.reduce_mean(tf.nn.sigmoid_cross_entropy_with_logits(output,Y))

  改为

  loss = tf.reduce_mean(tf.nn.sigmoid_cross_entropy_with_logits(labels=output,logits=Y))

• AttributeError: module ‘tensorflow’ has no attribute ‘batch_matmul’

  解决方法：batch_matmul() 改为 matmul()

• AttributeError: module ‘tensorflow’ has no attribute ‘pack’

  解决方法：pack() 改为 stack()

• TypeError: Expected int32, got list containing Tensors of type ‘_Message’ instead.

  解决方法：tf.cocat([fw,bw],2) 改为 tf.cocat(2,[fw,bw])

• Input ‘split_dim’ of ‘Split’ Op has type float32 that does not match expected type of int32

  解决方法：tf.split(axis, num_or_size_splits, value) 改为 tf.split(value, num_or_size_splits, axis)

• ValueError: Shapes (2, 100, 1) and () are incompatible

  解决方法：concated = tf.concat(1, [indices, sparse_labels]) 改为 concated = tf.concat([indices, sparse_labels], 1)

• AttributeError: module ‘tensorflow’ has no attribute ‘merge_all_summaries’

  解决方法：summary_op = tf.merge_all_summaries() 改为：summary_op = tf.summary.merge_all()

• AttributeError: ‘module’ object has no attribute ‘image_summary’

  解决方法：tf.image_summary() 改为 tf.summary.image()

• AttributeError: 'module’object has no attribute ‘SummaryWriter’

  解决方法：tf.train.SummaryWriter() 改为：tf.summary.FileWriter()

• AttributeError: ‘module’ object has no attribute ‘rnn_cell’

  解决方法：tf.nn.rnn_cell 改为 tf.contrib.rnn

• ValueError: Variable Wemb/Adam/ does not exist, or was not created with tf.get_variable(). Did you mean to set reuse=None in VarScope?

  解决方法：需要定义scope，报错是在optimizer处提示，但需要在定义模型时增加scope