grimoire/mmdetection-to-tensorrt
convert mmdetection model to tensorrt, support fp16, int8, batch input, dynamic shape etc.
PythonApache-2.0
Issues
- 4
Can't import mmdet2trt on jetson
#61 opened - 6
- 4
- 2
- 1
Support for detectron2
#57 opened - 2
- 0
cannot run demo/inference.py
#55 opened - 0
Problem running via docker
#54 opened - 3
- 5
- 2
DCNv2 support
#50 opened - 3
- 5
- 10
- 1
- 0
#assertion/root/space/amirstan_plugin/src/plugin/batchedNMSPlugin/batchedNMSPlugin.cpp,140
#45 opened - 0
kerError
#44 opened - 1
- 3
yolov3 can not use int8 calibration?
#42 opened - 4
int8 model convert error
#41 opened - 9
- 7
MemoryError on jetson TX2
#38 opened - 9
- 1
- 5
- 1
mmdet2trt run error
#34 opened - 6
- 9
- 0
- 5
- 3
Issue while converting with int8 support
#28 opened - 0
convert one-stage detector failed
#27 opened - 1
Mask rcnn can not use int8 calib?
#26 opened - 1
the question about interpolate
#25 opened - 7
GPU Memory
#24 opened - 12
请问转存出来的文件可以用C++调用吗?
#23 opened - 5
Some questions about this project
#22 opened - 1
Roadmap for instance segmentation models
#21 opened - 2
- 1
Bug: Generalized Focal Loss
#19 opened - 4
test.py
#18 opened - 3
- 7
error:batchedNMSPlugin.cpp
#16 opened - 1
- 2
skeleton or stgcn module
#13 opened - 1
教程
#12 opened - 1
GPU Memory-Usage Control
#11 opened - 2
run mmdet2trt() failed
#10 opened - 4
- 7
inference error
#7 opened