detectron2 faster rcnn to tensor rt

[Kolkhoznyk]

Hello is there any doable way to convert the detectron2 faster rcnn + fpn model to tensor rt engine? I found a tutorial and scripts about conversion of detectron2 mask rcnn model, i followed all the steps but it does not work with my architecture. Any suggestions? is it even worth trying? The goal is to detect small objects in complex scenes that's why i have chosen fr cnn. But it would be great to make the inference a bit faster if possible. Thanks

[poweiw]

it does not work with my architecture

There's many possible reasons that may lead to this. Could you provide your model for reference and maybe a brief description about what's different between your architecture and the mask rcnn? FYI there's also a faster rcnn sample at https://github.com/NVIDIA/TensorRT/tree/release/10.10/samples/python/dds_faster_rcnn, which uses a pre-trained faster rcnn from https://github.com/onnx/models/tree/refs/heads/main/validated/vision/object_detection_segmentation/faster-rcnn/model

[github-actions[bot]]

Issue has not received an update in over 14 days. Adding stale label. Please note the issue will be closed in 14 days after being marked stale if there is no update.

[github-actions[bot]]

This issue was closed because it has been 14 days without activity since it has been marked as stale.