RuntimeError: Not compiled with GPU support

[wangyansenhub]

Logging results to runs\train\exp35 Starting training for 300 epochs...

 Epoch   gpu_mem       box       obj       cls     theta    labels  img_size
 0/299      8.5G    0.1637   0.03505         0     10.14        33      1024: 100%|██████████| 39/39 [00:33<00:00,  1.15it/s]
           Class     Images     Labels          P          R  [email protected]  [email protected]:.95:   0%|          | 0/20 [00:00<?, ?it/s]

Traceback (most recent call last): File "train.py", line 640, in main(opt) File "train.py", line 537, in main train(opt.hyp, opt, device, callbacks) File "train.py", line 376, in train compute_loss=compute_loss) File "D:\ProgramData\envs\pytorch\lib\site-packages\torch\autograd\grad_mode.py", line 26, in decorate_context return func(*args, *kwargs) File "D:\Tian\yolov5_obb-master\val.py", line 206, in run out = non_max_suppression_obb(out, conf_thres, iou_thres, labels=lb, multi_label=True, agnostic=single_cls) # list(n, [xylsθ, conf, cls]) θ ∈ [-pi/2, pi/2) File "D:\Tian\yolov5_obb-master\utils\general.py", line 857, in non_max_suppression_obb _, i = obb_nms(rboxes, scores, iou_thres) File "D:\Tian\yolov5_obb-master\utils\nms_rotated\nms_rotated_wrapper.py", line 41, in obb_nms inds = nms_rotated_ext.nms_rotated(dets_th, scores, iou_thr) RuntimeError: Not compiled with GPU support 我总是在训练一定批次之后，报这个错误。都是在验证的时候中断，这是什么问题呢？

[elahe-badali]

Hi, I have the same issue. Any solution?

[wangyansenhub]

嗨，我有同样的问题。任何解决方案？

Not yet~

[hukaixuan19970627]

把这个步骤生成的.so文件删除，然后再编译一次。

[matteopantano]

Had the same issue @wangyansenhub , sometimes it is better to compile without cache, see my reply here #402

[wangyansenhub]

把这个步骤生成的.so文件删除，然后再编译一次。

LINK : fatal error LNK1181: 无法打开输入文件“D:\Tian\yolov5_obb-master\utils\nms_rotated\build\temp.win-amd64-cpython-39\Release\src\nms_rotated_cpu.obj” error: command 'C:\Program Files (x86)\Microsoft Visual Studio\2019\Community\VC\Tools\MSVC\14.29.30037\bin\HostX86\x64\link.exe' failed with exit code 1181 我重新安装以后 编译一直报错。这是什么问题呢? cuda：10.2 pytorch:1.10.0+cu102

[cheng1533]

把这个步骤生成的.so文件删除，然后再编译一次。

LINK : fatal error LNK1181: 无法打开输入文件“D:\Tian\yolov5_obb-master\utils\nms_rotated\build\temp.win-amd64-cpython-39\Release\src\nms_rotated_cpu.obj” error: command 'C:\Program Files (x86)\Microsoft Visual Studio\2019\Community\VC\Tools\MSVC\14.29.30037\bin\HostX86\x64\link.exe' failed with exit code 1181 我重新安装以后 编译一直报错。这是什么问题呢? cuda：10.2 pytorch:1.10.0+cu102

windows 下吧？可能是MSVC环境变量没有配好。

[haha457]

如果删除.so文件重新编译的方法不行，重新配置环境装pytorch，torchvision, torchaudio注意一定要与cuda版本对应

[elahe-badali]

I finally solve it by this solution: https://github.com/hukaixuan19970627/yolov5_obb/issues/402#issuecomment-1189014923