MLOpsManufacturing
MLOpsManufacturing copied to clipboard
liupeirong
Demonstrate samples and good engineering practice for operationalizing machine learning solutions.
Currently the business logic edge module uses while loop to run forever: ```python async def continuous_loop(): while True: time.sleep(100) ``` A good potential improvement for this edge module is to...
**IoT Edge Module Deployment: Find way to validate old edge modules have stopped before checking if new modules have started** Currently in our edgeSmokeTest we are sleeping after we validate...
**IoT Edge Integration Tests: use edgeHubDev instead of deploying VM for edge device sim** Currently in our integration test DevOps pipeline, we deploy a VM edge device in Azure to...
Currently, we are using pytest and unit test inside our business logic unit tests, refactor the tests to only use one framework
In most samples a new AML pipeline endpoint is created each time the pipeline has been (re)build. There is a more elegant way of updating the [existing endpoint](https://github.com/liupeirong/MLOpsManufacturing/blob/010e2cc9f06e49155b1ac87c85d54c9e0d8c0135/samples/kaldi-asr-yesno/ml_service/pipelines/build_pipeline.py#L64) in the...
Metadata
Owner
Metadata
Demonstrate samples and good engineering practice for operationalizing machine learning solutions.