mirrors/mmdeploy
1
0
Fork 0
mirror of https://github.com/open-mmlab/mmdeploy.git synced 2025-01-14 08:09:43 +08:00
Code Issues Projects Releases Wiki Activity
mmdeploy/docs/en/codebases/mmrotate.md
Yue Zhou 42dc5bc316
Support single stage rotated detector in MMRotate (#428) 
* fix lint

* fix lint

* add mmrotate part

* update

* update

* fix

* remove init_detector

* success run with bs=1

* nms_rotated support batch

* support [batch_id, class_id, box_id]

* fix

* fix

* Create test_mmrotate_core.py

* add ut

* add ut

* Update nms_rotated.py

* fix

* Revert "fix"

This reverts commit f792387fb449ba091c1d932f29d28214805fb6e3.

* add mmrotate into requirements

* add ut

* update doc

* update

* skip test because mmcv version < 1.4.6

* update

* Update rotated-detection_static.py

* Update rotated-detection_static.py

* Update rotated-detection_static.py

* fix bug of memory leak.

* Update rotated_detection_model.py
2022-05-07 16:11:43 +08:00

2.2 KiB
Raw Blame History

MMRotate Support

MMRotate is an open-source toolbox for rotated object detection based on PyTorch. It is a part of the OpenMMLab project.

MMRotate installation tutorial

Please refer to official installation guide to install the codebase.

MMRotate models support

Model Task ONNX Runtime TensorRT NCNN PPLNN OpenVINO Model config
RotatedRetinaNet RotatedDetection Y N N N N config

Example

# convert ort
python tools/deploy.py \
configs/mmrotate/rotated-detection_onnxruntime_dynamic.py \
$MMROTATE_DIR/configs/rotated_retinanet/rotated_retinanet_obb_r50_fpn_1x_dota_le135.py \
$MMROTATE_DIR/checkpoints/rotated_retinanet_obb_r50_fpn_1x_dota_le135-e4131166.pth \
$MMROTATE_DIR/demo/demo.jpg \
--work-dir work-dirs/mmrotate/rotated_retinanet/ort \
--device cpu

# compute metric
python tools/test.py \
    configs/mmrotate/rotated-detection_onnxruntime_dynamic.py \
    $MMROTATE_DIR/configs/rotated_retinanet/rotated_retinanet_obb_r50_fpn_1x_dota_le135.py \
    --model work-dirs/mmrotate/rotated_retinanet/ort/end2end.onnx \
    --metrics mAP

# generate submit file
python tools/test.py \
    configs/mmrotate/rotated-detection_onnxruntime_dynamic.py \
    $MMROTATE_DIR/configs/rotated_retinanet/rotated_retinanet_obb_r50_fpn_1x_dota_le135.py \
    --model work-dirs/mmrotate/rotated_retinanet/ort/end2end.onnx \
    --format-only \
    --metric-options submission_dir=work-dirs/mmrotate/rotated_retinanet/ort/Task1_results

Note

  • Usually, mmrotate models need some extra information for the input image, but we can't get it directly. So, when exporting the model, you can use $MMROTATE_DIR/demo/demo.jpg as input.

Reminder

None

FAQs

None