Skip to content

Latest commit

 

History

History
46 lines (23 loc) · 2.13 KB

step_five.md

File metadata and controls

46 lines (23 loc) · 2.13 KB
Raw

Saving and deploying the model

Saving the model

Once your training has completed and you have the model from training , now its time to save and deploy the model.

  1. You can save the model by clicking on the 3 vertical dots next to your model and click Save model

save_model_1

  1. Follow the steps and you will get a notification on the top saying Model successfully saved. View model details here You can follow the link.

save_model_3

  1. Alternatively, you can go to the assets tabs of your project to to view all your saved models.

save_model_4

Deploying and inferencing from the model

Once the model, the next step is to deploy or host the model.

  1. Go to the saved model and you will see the link to deploy the model Add deployment

deploy_model_1

  1. Follow the steps and you will get to the screen of deploying model. The status of deploying model will change from INITIALIZING to DEPLOY_SUCCESS

deploy_model_3

  1. Click on the name of the saved deployment and you will be directed to deployment details. You can click on the Implementation tab to get details on how to get to the model.

deploy_model_4

  1. The above gives details on how your model is hosted and how you can get to your model using a REST endpoint and/or how to query this model using different clients.

  2. You can also quickly test your model by using the Test tab. For the mnist example, copy the contents of this file which is a tensor representation of the number.

test_deploy_model_1

Further reading

  1. Overview: Deploying and scoring a deep learning model
  2. Deploying and scroing models using CLI