lmq5294249
lmq5294249
我将onnx模型转为NCNN模型后，再用NCNN来推理得出结果 [fast-reid_mobilenetv2-sim-opt.zip](https://github.com/linghu8812/yolov5_fastreid_deepsort_tensorrt/files/11284880/fast-reid_mobilenetv2-sim-opt.zip) fast-reid_mobilenetv2-sim-opt.param的推理网络图，可以看到输出blob的名称，不确定我选择的对不对   从结果可以获取的ncnn::mat 数据，如何解析这个结果，fastreid推理后的数据包含的特征点是怎么样的？
我尝试将模型转为onnx出现错误，无法解决
1、Export .pt reid models to: ONNX  2、Then I export to NCNN model [osnet_x0_75_msmt17_combineall-sim-opt.zip](https://github.com/KaiyangZhou/deep-person-reid/files/11343545/osnet_x0_75_msmt17_combineall-sim-opt.zip) 3、set up parameters const float m_mean_value[3] = { 0.485, 0.456, 0.406}; const float...
原项目地址https://github.com/KaiyangZhou/deep-person-reid/tree/master ## model | 模型 | モデル 1. original model [osnet_x0_75_msmt17_combineall.pth.zip](https://github.com/Tencent/ncnn/files/11351804/osnet_x0_75_msmt17_combineall.pth.zip) https://kaiyangzhou.github.io/deep-person-reid/MODEL_ZOO.html 2.我用作者提供的方法进行转换为ONNX模型 https://github.com/KaiyangZhou/deep-person-reid/tree/master/tools CMD为 python export.py -p "./osnet_x0_25_market1501.pth" -hp --imgsz 256 128 --include onnx 转换后为 [osnet_x0_75_msmt17_combineall.onnx.zip](https://github.com/Tencent/ncnn/files/11351820/osnet_x0_75_msmt17_combineall.onnx.zip) 这一步可以验证结果为正确 自己写的简单方法，应该是正确的吧!!!(边学边改) [detect.py.zip](https://github.com/Tencent/ncnn/files/11351828/detect.py.zip)...
我通过将模型pth转为onnx在转为ncnn模型，但是不知道怎么获取mean_vals，norm_vals参数 原项目地址：https://github.com/JDAI-CV/fast-reid/tree/6300bd756e341ca180c60b8e07c218b9eb40664a/tools/deploy Use official [fast-reid](https://github.com/JDAI-CV/fast-reid) to transform PyTorch weights to ONNX model. Run the following command 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 可以得到onnx模型： [fast-reid_mobilenetv2.onnx.zip](https://github.com/Tencent/ncnn/files/11267823/fast-reid_mobilenetv2.onnx.zip) 转换后的ncnn...