Tensorflow服务的模型失败，并显示以下错误：“您必须输入占位符张量的值”

Question

以下是所服务的张量流模型的signature def：

带有tag-set: 'serve'的MetaGraphDef包含以下内容：

SignatureDefs:

signature_def['serving_default']:
  The given SavedModel SignatureDef contains the following input(s):
    inputs['in'] tensor_info:
        dtype: DT_INT32
        shape: (-1, 10)
        name: input_sentences:0
  The given SavedModel SignatureDef contains the following output(s):
    outputs['out'] tensor_info:
        dtype: DT_INT32
        shape: (-1, 1)
        name: output_sentences:0
  Method name is: tensorflow/serving/predict

在为该模型提供服务之后，当我传递形状为[-1, 10]的输入时，出现以下错误：

"You must feed a value for placeholder tensor 'output_sentences' with dtype int32 and shape [?,1]"

即使output_sentences是输出的一部分。

请帮帮我。