Shulou(Shulou.com)06/02 Report--

How to deploy the TensorFlow model, many novices are not very clear about this, in order to help you solve this problem, the following editor will explain for you in detail, people with this need can come to learn, I hope you can gain something.

TensorFlow model deployment 1. Docker image pull # Download the TensorFlow Serving Docker image and repodocker pull tensorflow/serving2. Save model # set the version number field of the model when saving the model, otherwise the error model.save (". / data/models/zsh_test/1") 3 of the model version will not be found after deployment to docker. Deployment Model 3.1 single Model deployment # Mapping restapi port 8501 The path defined by the grpc port 850map model # defines the model name docker run-p 8501├── 8501\-v / path/to/model/models/zsh_test:/models/zsh_test/\-e MODEL_NAME=zsh_test-t tensorflow/serving# model structure ~ / PycharmProjects/pytorch-study/data/models ❯ tree. └── zsh_test └── 1 ├── assets model saved_model. Pb └── variables ├── variables.data-00000-of-00001 └── variables.index4 directories 3 files3.2 multiple model deployment

Directory structure

Save multiple models under the model saving folder models, and add models.config configuration files to configure the basic model information.

~ / PycharmProjects/pytorch-study/data ❯ tree. ├── exp_all_data.csv ├── exp_all_result.csv └── models ├── models.config ├── zsh_test │ └── 1 │ ├── assets │ ├── saved_model.pb │ └── variables │ ├── variables.data-00000-of-00001 │ └── variables.index └── zsh_test1 └── 1 ├── assets ├── saved_model.pb └── variables ├── variables.data-00000-of-00001 └── variables.index9 directories 9 files

Configuration file

Model_config_list: {config: {name: "zsh_test", base_path: "/ models/zsh_test", model_platform: "tensorflow"}, config: {name: "zsh_test1", base_path: "/ models/zsh_test1", model_platform: "tensorflow"}}

# deploy multiple models docker run-p 8501 tensorflow/serving 8501\-v / path/to/model/models/:/models/\-t model-- model_config_file=/models/models.config4. Test API # http://ip:prot/version/models/model_name # to view model information curl http://localhost:8501/v1/models/zsh_test response: {"model_version_status": [{"version": "1", "state": "AVAILABLE", "status": {"error_code": "OK" "error_message": ""}}]} # call the model curl-- location-- request POST 'http://localhost:8501/v1/models/zsh_test:predict'\-- header' Content-Type: application/json'\-- data'{"instances": [test_data]}'is it helpful for you to read the above? If you want to know more about the relevant knowledge or read more related articles, please follow the industry information channel, thank you for your support.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.