Datature Tutorials

Train a keypoint estimation model on Datature Nexus using YOLOv8. Annotate pose landmarks, configure training, and run inference in eight minutes.

Train an image classification model on Datature Nexus with YOLOv8. Upload images, assign labels, launch training, and evaluate results in five minutes.

Train an object detection model on Datature Nexus with YOLOv8. Annotate bounding boxes, configure training, and run predictions in eight minutes.

Train a video classification model on Datature Nexus using MoviNet. Upload video clips, assign labels, and build a working classifier in seven minutes.

Train models on your own GPU hardware using Datature Nexus. Install the runner agent, connect to your workspace, and launch jobs remotely.

Manage model artifacts, compare versions, export to TFLite or ONNX, and deploy inference API endpoints directly from Datature Nexus.

Use asset groups on Datature Nexus to split datasets into train, validation, and test sets. Organize images by source, condition, or batch.

Evaluate trained models on Datature Nexus. Review loss curves, mAP, precision, recall, and confusion matrices across multiple training runs.

Link Azure Blob Storage to Datature Nexus for direct dataset import. Configure credentials and pull training images without manual downloads.