如何在TensorFlow中恢复字典变量？

Question

我的模型中有一个字典变量

evaluation = {}
evaluation['accuracy'] = ...
evaluation['f1'] = ...
evaluation['precesion'] = ...
evaluation['recall'] = ....

训练并保存模型后，我想加载保存的元图并恢复变量。我想通过名称

从图中获取占位符

evaluation = graph.get_operation_by_name("evaluation").outputs[0]

然后可以使用它

evaluation = sess.run(evaluation, {input_x: ...})

我想我需要为我的变量设置名称，但是当我尝试

时

evaluation = tf.Variable({}, name="evaluation")

我收到以下错误

TypeError: Failed to convert object of type <type 'dict'> to Tensor. Contents: {}. Consider casting elements to a supported type.

Answer 1

tensorflow中没有字典变量。你的字典是在python中，并没有存储在张量流图中。如果您在加载图表后想要字典，则必须重新创建它：

evaluation = {}
evaluation["placeholder_name"] = graph.get_tensor_by_name('placeholder_name:0')