Retry a failed fine-tune
Requeues a failed fine-tune with fresh credentials. If a confirmed training job can be recovered, retry resumes monitoring that job and submits only the remaining epochs. Otherwise, if no remote training model exists, retry restarts preparation. Returns 409 Conflict if the training service reported the failure (remote_training_failed), the fine-tune is not failed, a previous training workflow is still running, or remote training cannot be recovered because its identity or confirmed submission history is unavailable. When restarting preparation, the source dataset must be ready at the original revision, and any selected relabel must be completed for that revision; otherwise the request returns 409 Conflict. These dataset and relabel checks do not apply when resuming confirmed training.
Authorizations
A W&B API key.
Headers
Accessible personal or team entity. Omit to use the authenticated user's default entity.
Response
The fine-tune was requeued.
0 <= x <= 2queued, training, deployed, failed 0 <= x <= 100Structured pipeline failure type. RemoteTrainingFailed means the training job cannot be resumed; create a new fine-tune instead of retrying.