Get startedOperate · Global feedback
Track and control tasks
Follow every long-running operation from one persistent workspace task center.
Source reviewed 2026-09-16. Availability depends on your installation, permissions, and compatible runtimes. A supported path is not a guarantee of model quality or production readiness.
Outcome
Leave a page without losing progress, then return directly to the running or completed operation.
Start here
What you need to know first
Check the inputs below before starting. If you are new, begin with the first-project guide. A dataset holds media and labels; a model produces results; a deployment makes a selected model version callable. Creating one does not create the others.
Bring these inputs
- — Any asynchronous platform operation
- — Task lifecycle events and source-page metadata
Expected output and limits
- — A durable state machine and progress history
- — Deep links to results or remediation
Core workflow
Launch any operation
Training, evaluation, labeling, conformity, deployment, and transfers create durable task state.
Follow progress
Select Runs in the bottom bar for quick progress, or Build & evaluate → Runs for the full run history. Each run shows its status, current phase, progress, and any failure context.
Control the task
Pause, resume, or cancel when the backend operation supports that transition.
Open the source
Select the task to return to its evidence. Platform operators can open Admin for an incident inventory that refreshes every 30 seconds and shows recovery actions; an unavailable refresh leaves the last result visibly stale.
What this surface supports
Active and recent tasks
Pause, resume, and cancel
Deep links
Failure and completion evidence
Expert section
Contracts, signals, and failure modes
Use this section when you are defining acceptance criteria, automating the surface, or reviewing whether its output is safe to promote downstream.
Quality and operating signals
- — Queued/running/paused/terminal state
- — Progress heartbeat
- — Retry count and terminal reason
Common failure modes
- — Stalled workers without heartbeat
- — Duplicate retries
- — A completed task with no discoverable artifact
Expert release checklist
□ Inputs and dependencies are pinned to immutable versions.
□ Acceptance metrics include critical classes and operating slices.
□ Failure, retry, cost, and rollback behavior are understood.
□ The resulting artifact has an owner and a downstream review path.
Engineering safeguards
- — State persists across navigation
- — Unsupported controls stay unavailable
- — Failures expose actionable context