数据说明路径：
https://github.com/JDAI-CV/fast-reid/tree/master/datasets
大概如下：

训练

./tools/train_net.py --config-file ./configs/Market1501/bagtricks_R50.yml MODEL.DEVICE "cuda:0"

评估：
python tools/train_net.py --config-file ./configs/Market1501/bagtricks_R50.yml --eval-only \
MODEL.WEIGHTS /path/to/checkpoint_file MODEL.DEVICE "cuda:0"

测试：
python tools/test_net.py --img_a1 a_1.jpg --img_a2 a_2.jpg --img_b1 b_1.jpg --img_b2 b_2.jpg --config-file ./configs/Market1501/bagtricks_R101-ibn.yml MODEL.WEIGHTS ../market_bot_R101-ibn.pth MODEL.DEVICE "cuda:6"

推理：
python tools/inference_net.py --img_a1 /data/disk1/workspace/05_DarknetSort/02_humanID/1.jpg --img_a2 /data/disk1/workspace/05_DarknetSort/02_humanID/21.jpg --img_b1 b_1.jpg --img_b2 b_2.jpg --config-file ../01_fast-reid/fast-reid/configs/Market1501/bagtricks_R101-ibn.yml MODEL.WEIGHTS ../market_bot_R101-ibn.pth MODEL.DEVICE "cuda:6"

转onnx
python tools/03_py2onnx.py --config-file ../fast-reid-master/configs/Market1501/bagtricks_R101-ibn.yml MODEL.WEIGHTS /data/disk1/project/model_dump/01_reid/market_bot_R101-ibn.pth MODEL.DEVICE "cuda:0"

python tools/03_py2onnx.py --config-file ./configs/Market1501/bagtricks_R101-ibn.yml MODEL.WEIGHTS ../market_bot_R101-ibn.pth MODEL.DEVICE "cuda:0"

转trt
python tools/03_check_onnx.py

./trtexec --onnx=/data/disk1/project/data/01_reid/02_changchuang/human_direction2.onnx --shapes=the_input:1x224x224x3 --workspace=4096 --saveEngine=/data/disk1/project/data/01_reid/02_changchuang/person_count_engine.trt

运行
python tools/04_trt_inference.py --model_path 4batch_fp16_True.trt

安装要求：以及所需求的包

Linux or macOS with python ≥ 3.6
PyTorch ≥ 1.6
torchvision that matches the Pytorch installation. You can install them together at pytorch.org to make sure of this.
yacs
Cython (optional to compile evaluation code)
tensorboard (needed for visualization): pip install tensorboard
gdown (for automatically downloading pre-train model)
sklearn
termcolor
tabulate
faiss pip install faiss-cpu