【torch.use_deterministic_algorithms运行报错解决办法】

东方yu晓

已于 2023-11-16 11:36:13 修改

阅读量535

点赞数 1

文章标签：深度学习 YOLO

于 2023-11-04 16:07:38 首次发布

本文链接：https://blog.csdn.net/qq_35900810/article/details/134219541

版权

YOLO训练时torch.use_deterministic_algorithms运行报错解决办法

- YOLO训练时出现如下错误
- 解决办法：在torch中(路径为：home/用户名/anaconda3/envs/py38/lib/python3.8/site-packages/torch)找到_init_.py，将以下Warn_only=True改为False.
- 完美解决，开始训练

YOLO训练时出现如下错误

// Error
RuntimeError: upsample_bilinear2d_backward_out_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True)'. You can turn off determinism just for this operation, or you can use the 'warn_only=True' option, if that's acceptable for your application. You can also file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation.

解决办法：在torch中(路径为：home/用户名/anaconda3/envs/py38/lib/python3.8/site-packages/torch)找到_init_.py，将以下Warn_only=True改为False.

def use_deterministic_algorithms(mode, *, warn_only=True):

完美解决，开始训练

poch    GPU_mem   box_loss   obj_loss   cls_loss  Instances       Size
      0/399      3.64G    0.07086    0.02843    0.02626         14        640: 100%|██████████| 361/361 01:09
                 Class     Images  Instances          P          R      mAP50   mAP50-95: 100%|██████████| 40/40 00:27
                   all        626        721   5.53e-05    0.00797   2.82e-05    5.8e-06