Upgrade to Pro — share decks privately, control downloads, hide ads and more …

Let’s start ML on GCP with AutoML

Let’s start ML on GCP with AutoML

sakajunquality

March 23, 2019
Tweet

More Decks by sakajunquality

Other Decks in Technology

Transcript

  1. - @sakajunquality - Jun Sakata - GDE, Cloud - SRE

    @ Ubie, inc. - Usually… - #GKE #Kubernetes #containers #DevOps etc. - Not ML Person... Who am I?
  2. AutoML - State-of-the-art performance - Get up and running fast

    - Generate high-quality training data (from the official website...)
  3. #kubernetes-docs-ja - A community translation project of Kubernetes Docs into

    Japanese - https://kubernetes.io/ja/docs/home/ - Slack - http://slack.k8s.io/ - #kubernetes-docs-ja channel
  4. Prepare the dataset - Use the already translated Japanese and

    original English Translated sentence pairs (en/ja) K8s specific translation Model
  5. Prepare the dataset - Use the already translated Japanese and

    original English - Some amendments - e.g. - Make 1:1 pairs of sentences - Use the same terms
  6. Prepare the dataset - Not enough sentences yet in the

    project - Use some sentences from GKE docs - https://cloud.google.com/kubernetes-engine/docs/ - Topic and terms are quite similar - With some amendments in terms
  7. Prepare the dataset - Export sentences pairs as TSV -

    And upload into AutoML Translation
  8. Train - Just Click “START TRAINING” - Can use base

    model - Google NMT (Default) - https://ai.google/research/pubs/pub45610 - Other AutoML model
  9. Result - Result is available with scores - Refer to

    “Evaluating Model” - https://cloud.google.com/translate/automl/docs/evaluate
  10. - Need more samples? - By default 10% of sentences

    are used for validation and test for each. - More datasets for training> - Model with datasets only from GKE is quite high in scores - Google’s official translation is better than community one? Considerations
  11. Takeaways - You can start ML easily with AutoML -

    Creating Model - Serving Model - Some updates in Next SF ‘19 ?