search

aws / sagemaker-tensorflow-serving-container

A TensorFlow Serving solution for use in SageMaker. This repo is now deprecated.
Apache License 2.0
172 stars 101 forks source link

Filtering by numeric model versions. #177

Closed samueleresca closed 4 years ago

samueleresca commented 4 years ago

Recently, I experienced an issue in deploying the sagemaker serving container. I was getting the following:

[libprotobuf ERROR external/com_google_protobuf/src/google/protobuf/text_format.cc:317] Error parsing text-format tensorflow.serving.ModelServerConfig: 8:19: Expected integer, got: .

The above error was caused by some hidden folders that were included in the targz process that creates the model. The issue has been introduced to the following PR: #155. I also see that there is already an open PR for excluding the code folder from the find_model_versions method (#160).

Instead of excluding explicitly the code folder I would prefer filtering by version.isnumeric():

".0005".isnumeric() -> False
"code".isnumeric() -> False
"0005".isnumeric() -> True
"5".isnumeric() -> True

I have included a hidden folder + nested file as part of the testing model.

By submitting this pull request, I confirm that my contribution is made under the terms of the Apache 2.0 license.

sagemaker-bot commented 4 years ago

AWS CodeBuild CI Report

Powered by github-codebuild-logs, available on the AWS Serverless Application Repository