Model Registy.

Once you've completed all the training processes and got your best model by analysing their performances. Then, you need to register the model. Registering your model is important step, because only then you can deploy your model.

Now you can deploy the model that you've registered from model registry tab.

• Click on the model that you've registered previously.

• Once you click on the registered model, you can see various details about the model. Like the model version, registered time, model's current stage and model training time.

Different stages for a model.

• None - Model is registered but not under staging or production.

• Staging - If the model is in Staging, the Quality Analysing team can check the performance of the model and do various types of test cases to measure the model's capabilities and scalability. If they're satisfied with the performance, they can move it to production.

• Production - If you want to send the model into the production environment you need to choose the Production option. Then it can be used through a model API or it can be used in Web Applications.

Note: You can deploy a model only when it is with the Production tag.

Once you tagged a model as Production staged, then you can use it in deployment.