Dataset.create_object_index( model_run_id: str, gt_only: bool ) -> AsyncJob
This endpoint will create or update an object index for a Dataset, provided that one exists.
This will generate internal embeddings for each object in the dataset that does not already have an associated embedding. These embeddings are used for object Autotag and similarity search.
This endpoint is idempotent. If this endpoint is called again for a model run that has been indexed in the past, the previously indexed predictions will not have new embeddings recomputed. The same is true for ground truth annotations. This means if you change update a prediction or ground truth bounding box that already has an associated embedding, the embedding will not be updated, even with another call to this endpoint. To simultaneously update an annotation/prediction and update its embedding, we recommend deleting the prediction or ground truth annotation and re-inserting it.
This endpoint is currently limited to indexing all the objects of a single model run, or all of the ground truth annotations of a dataset. It is also limited to index up to only 3 million objects at a time — if a request is made that exceeds this limit, the job will fail.
This endpoint is currently only enabled for enterprise customers. Please reach out to [email protected] if you wish to learn more.