​ ValueError: Input 0 of node ... was passed float from ... incompatible with expected float_ref.

将pb与ckpt通过freeze_graph.py整合生成新的pb文件，对文件执行图导入时，执行

tf.import_graph_def(graph_def, name='')

遇到如下报错：

ValueError: Input 0 of node lanenet_loss/inference/encode/conv1_1/cond/bn/AssignMovingAvg/Switch was passed float from lanenet_loss/inference/encode/conv1_1/bn/moving_mean:0 incompatible with expected float_ref.

原因是在运行freeze_graph.py之后，AssignSub和RefSwitch处理后的输入类型由float_ref变为了float，因此要额外地去转换一下,在图导入之前执行如下代码：

for node in graph_def.node :
	if node.op == 'RefSwitch':
		node.op = 'Switch'
		for index in range(len(node.input)):
			if 'moving_