Polar Bear AI

I have used my laptop(GPU is GTX1650) to run yolov7 and yolov5-l. At first, it seems yolov7(150ms/image) is slower than yolov5-l(70ms/image). But I found this [issue](https://github.com/WongKinYiu/yolov7/issues/175). When set `half=False`, the...

@why228430 @wangchengtza I met same problem on Tesla T4. And I have tried even 64 for **THREADS_PER_BLOCK** . Problem still exists. My env: GPU: Tesla T4 Pytorch: 1.3.1 with cudatoolkit=10.0...

Yes, I did re-compile "bash compile.sh". But problem still exists. Did my environment cause the error? I use docker to impl this project. docker: pytorch/pytorch:1.3-cuda10.1-cudnn7-devel I have installed ReDet following...

@why228430 Thanks a lot. I'll try new env. What is the training speed in your platform (2080Ti)?

@subzerofun Hello! Did you solve this problem? `ValueError: Shapes (64, 64, 64, 256) and (64, 32, 32, 256) are not compatible`

@mkocabas @mayoudong1993 This problem should not be from pyrender. The biggest problem is the code : os.environ['PYOPENGL_PLATFORM'] = 'egl' in [demo.py](https://github.com/mkocabas/VIBE/blob/cee26151ea880ea7547e78afd7faff92b102a353/demo.py#L18). It will use EGL as OpenGL Platform. So when...

> Thanks @StephenGreat for the solution. Is this the only modification you did to make VIBE work on Windows? You need to install ffmpeg (windows latest version) which can download...

@randi120 The eigen errors are solved. But I got "undefined symbol 'M_PI' in sophus.hpp". And I have compiled Sophus successfully. My platform: VS2015, CUDA8.0 Could you help me to solve...

I want to know whether the code is just for kinect (v1) ? And does it need to install kinect driver from MicroSoft or install 'Sensor openSource64'?