Importing model predictions
When running an importer, any previously imported predictions will be overwritten! To ensure the ability to revert to previous model iterations, it is important to version your projects.
If you aren't familiar with how to build lists of Prediction
objects, please have a look at the Import model predictions guide first.
It will show you how to construct predictions for bounding boxes, polygons, masks, and classifications.
With these predictions in hand, importing them is done as follows:
from pathlib import Path
from encord_active.lib.db.predictions import Prediction
from encord_active.lib.model_predictions.importers import import_predictions
from encord_active.lib.project import Project
project_path = Path("/path/to/your/project/root")
predictions: list[Prediction] = ... # Your list of predictions
import_predictions(Project(project_path), predictions)
Other Options
There are a couple of additional ways to import predictions that are stored in common file formats. You can find them here:
📄️ COCO
When running an importer, any previously imported predictions will be overwritten!
📄️ KITTI Bounding Boxes
The KITTI dataset is a widely-used computer vision dataset for benchmarking and evaluating autonomous driving systems.
📄️ Instance Segmentation Masks
The following approach only works for segmentation and polygons.