Deployments overview | Dagnam.AI Docs

What you can deploy today, how a deployment key works, and the statuses a deployment moves through.

Deployments overview A deployment serves one model version behind an API. This page covers what you can deploy today, how to deploy a version, the deployment key, and the statuses you will see. What you can deploy today You can deploy a ready version of a model you fine-tuned chat completion or text classification from My models . See Fine-tune a model /docs/training/fine-tune for how a version gets there. Access Deployments are not included in the Free plan, and serving a model needs cloud GPU access, available on request during the beta. Ask through Support /support . Deploy a model version When Deploy next to a version is disabled, a caption under it says why: Not servable yet when the version is not ready to deploy, or Not available on your plan. with a Contact Support link when your account cannot deploy today. --name ", , label: "Python", language: "python", code: 'deployment = dagnam.deployments.deploy model version \n version id,\n name="my-classifier v1",\n \napi key = deployment "api key" shown once', , / The deployment key A deployment key is separate from your personal API key: one key per deployment, sent as Authorization: Bearer on every inference call. It has no sk prefix, and it is shown in full only once, at creation and again if you regenerate it from the deployment's Settings tab. Statuses Status Meaning --------------- --------------------------------------- Not provisioned Created, not yet serving. Deploying Being set up. Running Serving requests. Paused Not accepting requests, by your choice. Failed Setup or serving failed. Health reads Healthy , Unhealthy , or Unknown ; it stays Unknown until the platform's own automatic check confirms the deployment is answering, not until you call it yourself. A deployment that was created by an earlier flow and never got a version deployed to it shows a Not served badge; the only action offered on it is Delete . Studio-trained models Deploying a model trained in Studio is Coming soon . The Deploy action on a dashboard project and on a training job's checkpoint is disabled, and the /deploy page points you to My models .