Torchscript jit scripting & Faster rCNN issue

Hello community,

I hope you’re all doing well. For the past few weeks, I’ve encountered a persistent issue that has me truly stumped, and I’m reaching out in the hopes that someone might have encountered this before or can shed some light.

Here’s the situation: I’ve exported a TorchScript model using (torchvision.models.detection.fasterrcnn_resnet50_fpn(pretrained=True)
However, every time I attempt to infer with this model, the output I receive is an empty list: [].

I was hopping to get a dictionary including boxes = instances, labels, scores.

Has anyone successfully read the output of a fasterrcnn model during TorchScript inference? I would be profoundly grateful for any guidance or suggestions.

Thank you in advance for considering my query. :pray: