search

jd730 / OICR-pytorch

Pytorch Implementation of Multiple Instance Detection Network with Online Instance Classifier Refinement
MIT License
75 stars 10 forks source link

Torch #27

Closed XuanX-new closed 3 years ago

XuanX-new commented 3 years ago

Sorry to bother you. I have the followed question. Found GPU0 GeForce RTX 2080 Ti which requires CUDA_VERSION >= 9000 for optimal performance and fast startup time, but your PyTorch was compiled with CUDA_VERSION 8000. My Configuration is torch 0.4.0+cuda9.0+torchvision0.2.2.

jd730 commented 3 years ago

Hi @xiexuan1117, the reason why I implemented on torch 0.40 is to use compiled roi_pool. So, I guess that you can use torchvision's roi pooling with higher version of torch.

XuanX-new commented 3 years ago
Thank you for your suggestion. I will try it. Have a nice weekend! 谢旋

xiexuan1117@163.com | 签名由网易邮箱大师定制 On 12/12/2020 09:08，Jaedong Hwangnotifications@github.com wrote：

Hi @xiexuan1117, the reason why torch version is 0.40 is to use compiled roi_pool. So, I guess that you can use torchvision's roi pooling with higher version of torch.

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or unsubscribe.

jd730 commented 3 years ago

FYI, please refer to https://github.com/jd730/OICR-pytorch/issues/8#issuecomment-653499183

XuanX-new commented 3 years ago

Sorry to bother you again. I have the follow issue. File "/home/summer/anaconda3/envs/OICR/lib/python3.6/site-packages/torch/nn/modules/conv.py", line 301, in forward self.padding, self.dilation, self.groups) RuntimeError: CUDNN_STATUS_EXECUTION_FAILED Do you know the reason for this？My Configuration is pytorch 0.4.0+cuda9.0+torchvision0.2.1.

jd730 commented 3 years ago

Hi @xiexuan1117, which GPU are you using? If I remember correctly, after GTX 20XX, Nvidia GPUs do not support torch 0.4

XuanX-new commented 3 years ago

ok，I see. The problem is that you say.The GPU is 2080 Ti that can‘t support torch 0.4. Thank you very much.