Wu Yan

VictorLamoine, Thanks for your reply. I have went to the gflag, glog, ceres directory to do`sudo make uninstall` and `make clean` , then re-compiling and installing as the url you...

@shaun-edwards Thank you for your infomation. Yes, I doubt the Ceres not compiled as a shared library, so I re-installed it again after add the PPA, but how can I...

https://forums.developer.nvidia.com/t/whats-the-proper-way-to-detect-sp-cuda-cores-count-per-sm/17657/10 ``` ~/NVIDIA_CUDA-10.2_Samples/1_Utilities/deviceQueryDrv$ ./deviceQueryDrv CUDA Device Query (Driver API) statically linked version Detected 1 CUDA Capable device(s) Device 0: "GeForce RTX 2060 SUPER" CUDA Driver Version: 11.1 CUDA Capability Major/Minor...

kinfu_remake/kfusion/src/cuda/temp_utils.hpp #if __CUDACC_VER_MAJOR__ >= 9 #include #endif ========================= list(APPEND CUDA_NVCC_FLAGS "-gencode;arch=compute_30,code=sm_30;-gencode;arch=compute_35,code=sm_35;-gencode;arch=compute_50,code=sm_50;-gencode;arch=compute_60,code=sm_70;-gencode;arch=compute_75,code=sm_75") ============================ __ballot(predicate) change to __ballot_sync(__activemask(), predicate); __all(predicate) change to __all_sync(__activemask(), predicate);

谢谢楼主，勾选bfloat16模式，可以下载更高精度的模型，但是打开以前用float32标注的文件夹就报错了，是不兼容是吧，是不是只能重新标注了？ 

我这边遇到了，在dump(coco_anno, f) 之前调用coco_anno = convert_numpy_types(coco_anno)确实解决了 

Could you provide the original .onnx model file in the project, so I can convert it myself. System Information: GPU Model (NVIDIA RTX 4070Ti S ) CUDA Version (11.8) cuDNN...

use glia_trt from here: http://rjyfb:[email protected]:10801/simple/glia-trt/