What are Inference Batch Jobs?

Inference batch jobs allow for automation of inference jobs on large datasets. Instead of calling our Inference API multiple times in a sequential manner, batch jobs operate on a dataset and leverage scaling and optimizations to efficiently process predictions on large numbers of images and videos.

Batch jobs can be integrated with webhooks that automatically send prediction results to a specified public HTTPs endpoint. Future integration expansions will also allow for the scheduling of cron jobs, setting up of notification and alert systems based on certain batch job events etc.

The main function categories are as follows:

Function Category	Description
Webhook Management	This governs the creation and management of webhooks that can integrated with batch jobs to automatically send prediction results to an endpoint URL.
Dataset Management	This governs the creation and management of datasets that the batch jobs will run inference on.
Job Management	This governs the creation, scheduling and management of batch jobs for optimized and efficient inference on large datasets.