Manage deployments

Deep Learning studio supports deployment of models.

How to deploy a model

You can deploy a model that you trained in deep learning studio and you can use it repeatedly.No need to deploy every time to use it.The deployed link will show up in "Deployment" tab.

Deployment Instructions

Use Deep Learning Studio project tab to use deployment feature.

1) Choose one sample project and copy it by clicking on copy icon

2) Open the project by clicking on "Project" tab, then switch to "Training" tab and start training

3) Switch to "Inference/Deploy" tab

4) Choose run, and then name the deployment,select on which instance you want to deploy on -

i) This instance

ii) Remote instance

5) Click on "Deploy"

i) This Instance (On your local instance) :

  • Select the “Training Run” and put any service name and click on “Deploy”.

  • Then you will get a link to do inference. Click on that link, it will redirect to deployment page. There you will see 2 feature i.e.,

a) Webapp :- Here, upload zip file or image to do inference and see the results.

b) API :- Run the curl command by running it to your local computer using console.

ii) Remote Instance :

  • Select the run and enter the service name, choose “Remote Instance”.

  • Enter SSH server name/IP, Port Number, Username and password.

  • Click on Deploy.

  • Then you will see the link, click on that link, and check the deployment.