服务但张量类型不匹配的输出模型

Question

通过在网上搜索您的问题，您找到了哪些相关的GitHub问题或StackOverflow线程？

我正在尝试导出服务模型，但是关于输入张量的报告类型错误。 但在导出和预测部分，输入是相同的类型。

如果可能的话，提供一个可重复性最小的示例（我们通常没有时间阅读数百行代码）

这是我导出的示例代码 ```

    named_graph_signature = {
        'inputs': exporter.generic_signature({
                      'sparse_index': tf.placeholder(tf.int64, name="feature_index")
                      'sparse_ids': tf.placeholder(tf.int64,name = "feature_ids"),
                      'sparse_values':tf.placeholder(tf.int64, name ="feature_values"),
                      'sparse_shape':tf.placeholder(tf.int64, name="feature_shape")
        }),
        'outputs': exporter.generic_signature({
            'prob': inference_softmax
        })}
    model_exporter.init(
        sess.graph.as_graph_def(),
        #default_graph_signature=named_graph_signature,
        named_graph_signatures=named_graph_signature,
        init_op=init_op)
    model_exporter.export(export_path, tf.constant(export_version), sess)
    print('Done exporting!')

```

这是我的预测代码

```

  ins = "0 142635:1 250810:1 335229:1 375278:1 392970:1 506983:1 554566:1 631968:1 647823:1 658803:1 733446:1 856305:1 868202:1"
  FEATURE_SIZE = 1000000
  tokens = ins.split(" ")
  feature_num = 0
  feature_ids = []
  feature_values = []
  feature_index = []

  for feature in tokens[1:]:
      feature_id, feature_value = feature.split(":")
      feature_ids.append(int(feature_id))
      feature_values.append(float(feature_value))
      feature_index.append([1, feature_num])
      feature_num += 1

  feature_shape = [1, FEATURE_SIZE]

  sparse_index = tf.contrib.util.make_tensor_proto(numpy.asarray(feature_index), dtype=tf.int64)
  sparse_ids = tf.contrib.util.make_tensor_proto(numpy.asarray(feature_ids), dtype=tf.int64)
  sparse_values = tf.contrib.util.make_tensor_proto(numpy.asarray(feature_values), dtype=tf.float32)
  sparse_shape= tf.contrib.util.make_tensor_proto(numpy.asarray(feature_shape), dtype=tf.int64)

  channel = implementations.insecure_channel(host, port)
  stub = prediction_service_pb2.beta_create_PredictionService_stub(channel)
  request = predict_pb2.PredictRequest()
  request.model_spec.name = model_name
  request.model_spec.version.value = model_version
  print model_name,model_version

  request.inputs['sparse_index'].CopyFrom(sparse_index)
  request.inputs['sparse_ids'].CopyFrom(sparse_ids)
  request.inputs['sparse_values'].CopyFrom(sparse_values)
  request.inputs['sparse_shape'].CopyFrom(sparse_shape)
  # Send request

  result = stub.Predict(request, request_timeout)

```

有用的日志或其他输出

（如果日志很大，请上传为附件或提供链接）。

```

Traceback (most recent call last):
  File "run.py", line 63, in <module>
    main()
  File "run.py", line 59, in main
    result = stub.Predict(request, request_timeout)
  File "/home/serving/.local/lib/python2.7/site-packages/grpc/beta/_client_adaptations.py", line 305, in __call__
    self._request_serializer, self._response_deserializer)
  File "/home/serving/.local/lib/python2.7/site-packages/grpc/beta/_client_adaptations.py", line 203, in _blocking_unary_unary
    raise _abortion_error(rpc_error_call)
grpc.framework.interfaces.face.face.AbortionError: AbortionError(code=StatusCode.INVALID_ARGUMENT, details="input size does not match signature")

```