search

yanxp / MetaR-CNN

Meta R-CNN : Towards General Solver for Instance-level Low-shot Learning
https://yanxp.github.io/metarcnn.html
177 stars 23 forks source link

RuntimeError: CUDNN_STATUS_EXECUTION_FAILED #52

Open NHW2017 opened 3 years ago

NHW2017 commented 3 years ago

Maybe you should update the version. The torch and CUDA used in this version are too old, and there are many problems in the implementation.

renqiqi commented 3 years ago

Maybe you should update the version. The torch and CUDA used in this version are too old, and there are many problems in the implementation.

have you solved this problem?

NHW2017 commented 3 years ago

Maybe you should update the version. The torch and CUDA used in this version are too old, and there are many problems in the implementation.

have you solved this problem?

No, I haven't solved this problem yet, if you run this program successfully, I hope you can give me a little help

NHW2017 commented 3 years ago

Maybe you should update the version. The torch and CUDA used in this version are too old, and there are many problems in the implementation.

have you solved this problem?

This week, I finally ran the code successfully. My previous environment was Ubuntu 16.04/18.04 rtx2080ti cuda8.0 / 9.0 / 9.2. Yes, I almost tried it, but failed. I didn't list too many problems. This week, I ran the code easily under Ubuntu 18.04 gtx1070 cuda9.0, and then I tried ubuntu18.04 1050ti / 1080ti cuda9.0 / 9.2, all of which were successful. However, the previous problems still exist in 2080ti and 2070s

renqiqi commented 3 years ago

Maybe you should update the version. The torch and CUDA used in this version are too old, and there are many problems in the implementation.

have you solved this problem?

This week, I finally ran the code successfully. My previous environment was Ubuntu 16.04/18.04 rtx2080ti cuda8.0 / 9.0 / 9.2. Yes, I almost tried it, but failed. I didn't list too many problems. This week, I ran the code easily under Ubuntu 18.04 gtx1070 cuda9.0, and then I tried ubuntu18.04 1050ti / 1080ti cuda9.0 / 9.2, all of which were successful. However, the previous problems still exist in 2080ti and 2070s

thank you! After that, I did not continue the experiment.I will try it according to your tips.

pengyinxw commented 3 years ago

Maybe you should update the version. The torch and CUDA used in this version are too old, and there are many problems in the implementation.

have you solved this problem?

This week, I finally ran the code successfully. My previous environment was Ubuntu 16.04/18.04 rtx2080ti cuda8.0 / 9.0 / 9.2. Yes, I almost tried it, but failed. I didn't list too many problems. This week, I ran the code easily under Ubuntu 18.04 gtx1070 cuda9.0, and then I tried ubuntu18.04 1050ti / 1080ti cuda9.0 / 9.2, all of which were successful. However, the previous problems still exist in 2080ti and 2070s

Hi @NHW2017, Did you following the other experiment requirements as the author suggested, except the cuda version? Thank you!