Skip to main content

Low Code Web UI

Low Code Web UI is a web-based user interface that allows users to manage the model cards, deploy models to GPU Cloud and on-premise devices, as well as monitor the endpoint status.

Register an Nexus AI Platform Account​

Before you start, you will need to create an account on FedML Nexus AI Cloud. After you create an account, you will see an Account Key (API Key) from the profile page. getApiKey.jpg

Login to Nexus AI Platform​

To login to Nexus AI Platform, you need to use fedml login $api_key command. Replace $api_key with your own API key.

fedml login $api_key

Create a Model Card Locally​

Use fedml model create command to create a model card locally. In this example, we will create a model card for EleutherAI/pythia-70m.

fedml model create -n hf_pythia_70m -m hf:EleutherAI/pythia-70m

Push and Check the Model Card on Nexus AI Platform​

Use fedml model push command to push a local model card to Nexus AI Platform.

fedml model push -n hf_pythia_70m

After that, you can check the model card uploaded to Nexus AI Platform. modelCardsUI.jpg

Deploy to GPU Cloud using UI​

You can deploy the model card to GPU Cloud using the "Deploy" button on the UI.

modelCardsDeployButton.jpg

After you click the button, you will be redirected to the deployment page. To deploy the model card to GPU Cloud, you need to first give a name to the endpoint. Then in the "Computing Source" section, select "FedML Cloud". After that, you can click the "Deploy" button to deploy the model card.

CreateEndpointGPUCloud.jpg

Deploy to on-premise devices using UI​

We will use the same model card called hf_pythia_70m that has been pushed to Nexus AI Platform in the previous section.

Bind your devices to Nexus AI Platform​

Before on-premise deploy, you need to bind your device to Nexus AI Platform.

fedml device bind $api_key

Check your device id on Nexus AI Platform (In our example is 32314).
OnPremDevices.jpg

modelCardsDeployButton.jpg

After you click the button, you will be redirected to the deployment page. To deploy the model card to GPU Cloud, you need to first give a name to the endpoint. Then in the "Computing Source" section, select "On-Premise". Then select the device id for master and worker. After that, you can click the "Deploy" button to deploy the model card.

onPremiseDeployUI.jpg