Trainings
Trainings
The run history — every training is scored, immutable, and deployable.
A training is one run of a pipeline: the platform takes the pipeline's segment data, trains the model, evaluates it, and writes a scored record. Trainings are immutable history — you can list them, inspect any run's full report, and deploy any completed run, not just the latest one.
Thirty seconds of API
# The latest runs across the workspace
curl "$API_BASE/v1/training?per_page=10" \
-H "Authorization: Bearer $TOKEN" \
-H "X-Workspace-Id: $WORKSPACE_ID"
# One run's full report
curl "$API_BASE/v1/training/$TRAINING_ID" \
-H "Authorization: Bearer $TOKEN" \
-H "X-Workspace-Id: $WORKSPACE_ID"Those are the only two dedicated training routes — trainings are started elsewhere:
| Trigger | Route |
|---|---|
| On demand from a pipeline | POST /v1/pipelines/{pipeline_id}/train |
| On the pipeline's schedule | training_schedule on the pipeline |
| From a goal | POST /v1/goals/{goal_id}/train |
In this section
Concepts
Lifecycle, the report, champions, and what a run costs.
Browse runs
List runs with filters and pagination; inspect one run's full report.
Endpoints
Every training route at a glance.
Putting a completed run live is Deployments; running forecasts against it is Inference.