liupeirong / MLOpsManufacturing

Demonstrate samples and good engineering practice for operationalizing machine learning solutions.
Other
21 stars 9 forks source link

Apply AML pipeline endpoint update (vs recreation) to more samples #74

Open h2floh opened 3 years ago

h2floh commented 3 years ago

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 in the Kaldi example.

I think we should update the image-classification-tensorflow and the non-python-preprocess samples with that pattern.