Assets and datasets
Import images, video or cloud assets, organize source data by project, and prepare samples for annotation and training.
Import → Organize → Select samplesPrepare your data.
Train the next model.
Organize assets, annotations, dataset versions and training tasks in one project. Use AI to prepare data while keeping human review in the workflow.
Keep the training work together.
- Assets and datasets
- Assisted annotation and review
- Published dataset versions
- Training, evaluation and export
Work on the data and the model.
These features reflect the platform’s current project and task workflows. Models, resources and interfaces depend on project configuration.
AI-assisted annotation
Generate draft annotations with AI, refine rectangles and polygons in the editor, then review before publishing a dataset version.
Draft labels → Human review → PublishImage generation
Prepare generated assets for a task and review candidate relevance. Keep generated samples separate from real test data.
Task description → Candidates → SelectionTraining and evaluation
Choose a published dataset version and base model, create a CV or VLM training task, inspect evaluation results and export the model.
Version → Train → Evaluate → ExportKeep review and traceability in the workflow.
Assisted annotation produces candidates that need review before publication. Training does not guarantee an improvement every time; compare results on independent test data when selecting a model version.
See the training workflowOpen your training workspace
Start with your data and your task.
Access is limited to invited test accounts. Sign in to an empty personal workspace; data pipelines and training execution are not connected yet.
Product enquiries
Tell us which product you need, your current equipment and the problem you want to solve.