bevdet-tensorrt-cpp
bevdet-tensorrt-cpp copied to clipboard
LCH1238
BEVDet implemented by TensorRT, C++； Achieving real-time performance on Orin
Thank you for this great work! I used the bevdet-export repo to convert 'bevdet-lt-d-ft-nearest.pth' or 'bevdet-r50-4dlongterm-depth-cbgs.pth' to onnx format, but got an error as follow: "./BEVDet-export/mmdet3d/models/detectors/trt_model.py", line 67, in forward...
您好，感谢您的工作。我在使用one分支测试bevdet.yaml，运行时YAML:会报错，我看了代码发现是bevdet.yaml中有些参数没有给出来，我测试bevdet_lt_depth.yaml配置时，使用tools/convert_bevdet_to_TRT.py脚本将pth转engine时，会出现KeyError: 'BEVDepth4DTRT is not in the models registry'错误，转模型使用的分支是BEVDetv2.1。我应该怎么处理才能使bevdet-r50-4dlongterm-depth-cbgs.pth文件转成engine文件？谢谢
使用[https://github.com/LCH1238/BEVDet/tree/export中的export_onnx.py，权重使用https://drive.google.com/drive/folders/1jSGT0PhKOmW3fibp6fvlJ7EY6mIBVv6i路径中的bevdet-lt-d-ft-nearest.pth导出onnx，最终的img_stage_lt_d.onnx如下图所示，与作者您提供的img_stage_lt_d.onnx不一致](https://github.com/LCH1238/BEVDet/tree/export%E4%B8%AD%E7%9A%84export_onnx.py%EF%BC%8C%E6%9D%83%E9%87%8D%E4%BD%BF%E7%94%A8https://drive.google.com/drive/folders/1jSGT0PhKOmW3fibp6fvlJ7EY6mIBVv6i%E8%B7%AF%E5%BE%84%E4%B8%AD%E7%9A%84bevdet-lt-d-ft-nearest.pth%E5%AF%BC%E5%87%BAonnx%EF%BC%8C%E6%9C%80%E7%BB%88%E7%9A%84img_stage_lt_d.onnx%E5%A6%82%E4%B8%8B%E5%9B%BE%E6%89%80%E7%A4%BA%EF%BC%8C%E4%B8%8E%E4%BD%9C%E8%80%85%E6%82%A8%E6%8F%90%E4%BE%9B%E7%9A%84img_stage_lt_d.onnx%E4%B8%8D%E4%B8%80%E8%87%B4) 
Hey, How to do quantization to int8 model of img_stage_lt_d.onnx and bev_stage_lt_d.onnx? Thanks!
fp32和fp16基本一致，检测框在120个左右 int8，检测框只能在10个左右了 其他啥都没动，仅仅是orin板端转成了int8的engine，用的--best
img_input = torch.zeros([6*B, 3, 900, 400], dtype=torch.int32, device=f'cuda:{args.gpu_id}') 请问在转onnx时为什么要设置成900*400，不直接设置为900*1600？
你好，标定参数文件sample0000.yaml 下的ego2global旋转平移参数有一个全局的，每个相机参数下面也有同样的参数，看了下每个相机下的旋转平移参数与全局下的ego2global_rotation ego2global_translation 差别不大，但还是有差别。请问这每个相机下的这两组参数与全局下的这两组参数有区别吗？谢谢
Thanks for such a great work. I want to do inference using camera. So , can you suggest changes that need to be done for that or can you add...
Hi: Thanks for this great work. I have successfully generate sample0_lidarbox.txt, then how to project perception results into images? best regards
