Обслуживание Tensorflow и сбой вызова клиента

Question

Что-то идет не так, потому что я получаю сбой соединения при попытке сделать клиентский вызов с помощью (localhost: 8517 или 10.110.110.13:8517)

root@ubuntu-xenial:~/serving# docker run -p 8517:8517 --mount type=bind,source=/root/serving/Ser_Model,target=/models/firstmodel -e MODEL_NAME=firstmodel -t tensorflow/serving
2018-12-18 22:42:54.202137: I tensorflow_serving/model_servers/server.cc:82] Building single TensorFlow model file config:  model_name: firstmodel model_base_path: /models/firstmodel
2018-12-18 22:42:54.203105: I tensorflow_serving/model_servers/server_core.cc:461] Adding/updating models.
2018-12-18 22:42:54.203162: I tensorflow_serving/model_servers/server_core.cc:558]  (Re-)adding model: firstmodel
2018-12-18 22:42:54.305763: I tensorflow_serving/core/basic_manager.cc:739] Successfully reserved resources to load servable {name: firstmodel version: 1}
2018-12-18 22:42:54.305814: I tensorflow_serving/core/loader_harness.cc:66] Approving load for servable version {name: firstmodel version: 1}
2018-12-18 22:42:54.305856: I tensorflow_serving/core/loader_harness.cc:74] Loading servable version {name: firstmodel version: 1}
2018-12-18 22:42:54.305899: I external/org_tensorflow/tensorflow/contrib/session_bundle/bundle_shim.cc:363] Attempting to load native SavedModelBundle in bundle-shim from: /models/firstmodel/1
2018-12-18 22:42:54.306021: I external/org_tensorflow/tensorflow/cc/saved_model/reader.cc:31] Reading SavedModel from: /models/firstmodel/1
2018-12-18 22:42:54.313227: I external/org_tensorflow/tensorflow/cc/saved_model/reader.cc:54] Reading meta graph with tags { serve }
2018-12-18 22:42:54.314376: I external/org_tensorflow/tensorflow/core/platform/cpu_feature_guard.cc:141] Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX2
2018-12-18 22:42:54.341301: I external/org_tensorflow/tensorflow/cc/saved_model/loader.cc:162] Restoring SavedModel bundle.
2018-12-18 22:42:54.354011: I external/org_tensorflow/tensorflow/cc/saved_model/loader.cc:138] Running MainOp with key legacy_init_op on SavedModel bundle.
2018-12-18 22:42:54.354072: I external/org_tensorflow/tensorflow/cc/saved_model/loader.cc:259] SavedModel load for tags { serve }; Status: success. Took 48045 microseconds.
2018-12-18 22:42:54.354116: I tensorflow_serving/servables/tensorflow/saved_model_warmup.cc:83] No warmup data file found at /models/firstmodel/1/assets.extra/tf_serving_warmup_requests
2018-12-18 22:42:54.354305: I tensorflow_serving/core/loader_harness.cc:86] Successfully loaded servable version {name: firstmodel version: 1}
2018-12-18 22:42:54.356224: I tensorflow_serving/model_servers/server.cc:286] Running gRPC ModelServer at 0.0.0.0:8500 ...
[warn] getaddrinfo: address family for nodename not supported
2018-12-18 22:42:54.356982: I tensorflow_serving/model_servers/server.cc:302] Exporting HTTP/REST API at:localhost:8501 ...
[evhttp_server.cc : 237] RAW: Entering the event loop ...

Попытка получить статусМодель выглядит следующим образом:

$ curl http://localhost:8517/v1/models/firstmodel

Но я получил сообщение об ошибке curl: (56) Recv failure: Connection reset by peer

Не могли бы вы выяснить, что не так в моей команде подачи через докер?

Answer 1

8517 - это порт, который открывает ваш докер-контейнер.Но вы запустили свой tf-сервер через порт 8500 и 8501 в контейнере.Если вы хотите, чтобы это работало, измените ваш docker cmd на docker run -p 8501:8501 xxx и измените URL запроса на localhost:8501/xxx/xxx.

Answer 2

Использование порта (8500 для GRPC и 8501 для RESTful API) имеет большое значение для Решение

Обслуживание для клиентов GRPC

docker run -p 8500:8500 --mount type=bind,source=/root/serving/Ser_Model,target=/models/firstmodel -e MODEL_NAME=firstmodel -t tensorflow/serving

ОбслуживаниеКлиенты RESTful API

 docker run -p 8501:8501 --mount type=bind,source=/root/serving/Ser_Model,target=/models/firstmodel -e MODEL_NAME=firstmodel -t tensorflow/serving

А также вы можете проверить статус RESTful API-клиента и вызовов,

curl http://localhost:8501/v1/models/firstmodel

curl -d '{"signature_name":"predict_output","instances":[2.0,9.27]}' -X POST http://10.110.110.13:8501/v1/models/firstmodel:predict