Question

tensorflow-gpu 1.10.0
tensorflow服务器1.10.0

我已经部署了一个Tensorflow服务器，该服务器可以提供多种模型。客户端代码就像client.py这样，我将其称为预测函数。

channel = implementations.insecure_channel(host, port)
stub = prediction_service_pb2.beta_create_PredictionService_stub(channel)
request = predict_pb2.PredictRequest()

def predict(data, shape, model_name, signature_name="predict"):
    request.model_spec.name = model_name
    request.model_spec.signature_name = signature_name
    request.inputs['image'].CopyFrom(tf.contrib.util.make_tensor_proto(data, shape=shape))
    result = stub.Predict(request, 10.0)
    return result.outputs['prediction'].float_val[0]

我有大约100个具有相同配置的客户端。以下是调用predict函数的示例代码：

from client import predict
while True:
    print(predict(data, shape, model_name))
    # time.sleep some while

首先，当我运行客户端代码时，我可以正确接收响应。但是几个小时后，客户端因错误而崩溃

_Rendezvous of RPC that terminated with (StatusCode.UNAVAILABLE, Socket closed)

我尝试将客户代码修改为

def predict(data, shape, model_name, signature_name="predict"):
    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.signature_name = signature_name
    request.inputs['image'].CopyFrom(tf.contrib.util.make_tensor_proto(data, shape=shape))
    result = stub.Predict(request, 10.0)
    return result.outputs['prediction'].float_val[0]

这意味着每次调用predict函数时，我都会尝试与tfs服务器建立连接。但是此代码也像以前一样失败。

那我应该怎么处理呢？

Answer 1

最后，我在channel.close()之前添加了return，它工作正常。

以（StatusCode.UNAVAILABLE，套接字已关闭）结尾的RPC的集合点

1 个答案: