Get startedMove labels and multimodal training evidence between Score and external systems.
Source reviewed 2026-09-23. Availability depends on your installation, permissions, and compatible runtimes. A supported path is not a guarantee of model quality or production readiness.
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
- A version-scoped annotation package
- File identity and class/schema mappings
Expected output and limits
- Imported evidence or a portable export archive
- Validation and rejection report
Core workflow
- 01
Choose a version
Open the exact dataset version whose schema and evidence should move.
- 02
Select a format
Choose a compatible specialist or multimodal representation and inspect schema mapping.
- 03
Validate the preview
Review matched files, classes, regions, prompts, answers, captions, and rejected records.
- 04
Run the transfer
Track the durable task and retain the generated report or export artifact.
What this surface supports
- Detection and segmentation formats
- VLM prompts, answers, and captions
- Linked text spans and regions
- Transfer summaries
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.
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.