fast-reid

fast-reid PyTorch=>ONNX=>TensorRT

1.Reference

• fast-reid github: https://github.com/JDAI-CV/fast-reid
• fast-reid arxiv: FastReID: A Pytorch Toolbox for General Instance Re-identification

2.Export ONNX Model

git clone https://github.com/JDAI-CV/fast-reid.git
cd fast-reid/tools/deploy
python onnx_export.py --config-file root-path/bagtricks_R50/config.yml --name baseline_R50 --output outputs/onnx_model --opts MODEL.WEIGHTS root-path/logs/market1501/bagtricks_R50/model_final.pth

3.Build tensorrt_inference Project

cd ../  # in project directory
mkdir build && cd build
cmake ..
make -j

4.run tensorrt_inference

cd ../../bin/
./tensorrt_inference fastreid ../configs/fast-reid/config.yaml ../samples/pedestrian_reid

results:

result
	1	0.93663549	0.63691914	0.65447605	0.59582067	0.63569337	0.67968875	0.6704126	0.64113915	0.6652534
	0.93663549	1	0.60606658	0.63805777	0.58507961	0.62270737	0.65357053	0.64495277	0.65284139	0.67049634
	0.63691914	0.60606658	1	0.85554099	0.64959818	0.65541494	0.63463104	0.58980453	0.64554667	0.63664776
	0.65447605	0.63805777	0.85554099	1	0.65920269	0.64546573	0.64599162	0.59834015	0.66127765	0.65907228
	0.59582067	0.58507961	0.64959818	0.65920269	1	0.87445605	0.72793484	0.7426371	0.68983912	0.69091672
	0.63569337	0.62270737	0.65541494	0.64546573	0.87445605	1	0.77302706	0.77868611	0.69440424	0.70926988
	0.67968875	0.65357053	0.63463104	0.64599162	0.72793484	0.77302706	1	0.91233504	0.65651429	0.66451794
	0.6704126	0.64495277	0.58980453	0.59834015	0.7426371	0.77868611	0.91233504	1	0.63502818	0.64097649
	0.64113915	0.65284139	0.64554667	0.66127765	0.68983912	0.69440424	0.65651429	0.63502818	1	0.92471766
	0.6652534	0.67049634	0.63664776	0.65907228	0.69091672	0.70926988	0.66451794	0.64097649	0.92471766	1