Parallel inference with visual genome categories.

[mbajaj01]

I want to do inference on big dataset of images. tools/infer_simply.py does inference by processing each image sequentially. All I want is to extract bounding boxes and their features. Can someone please guide me on how to do this more efficiently. Is the support already available to run inference on batches rather than doing for each image individually? I am new to the detectron framework. Any help is really appreciated.