search

sshaoshuai / PointRCNN

PointRCNN: 3D Object Proposal Generation and Detection from Point Cloud, CVPR 2019.
MIT License
1.7k stars 425 forks source link

Segmentation error #186

Open uranuscc opened 3 years ago

uranuscc commented 3 years ago

I am using linux RHEL6, python 3.7, pytorch 1.0. I was following the quick demo by running: $python eval_rcnn.py --cfg_file cfgs/default.yaml --ckpt PointRCNN.pth --batch_size 1 --eval_mode rcnn --set RPN.LOC_XZ_FINE False

when the code is executing rpn inference "rpn_output = self.rpn(input_data)" it reports segmentation fault. The input data looks fine. Does anyone have idea for this error, is it something wrong with c++ pointer?

Thanks! CC

K-Young96 commented 3 years ago

I also have this problem, but I don't know how to solve it

uranuscc commented 3 years ago

I have tested it on Ubuntu v20.04 with python 3.8, pytorch 1.0. The demo run works fine now. I think it's OS env issue with my RHEL6. cuda version or OS. some incompatibility there. still troubleshooting it...

QLuanWilliamed commented 3 years ago

I have tested it on Ubuntu v20.04 with python 3.8, pytorch 1.0. The demo run works fine now. I think it's OS env issue with my RHEL6. cuda version or OS. some incompatibility there. still troubleshooting it...

Hi, I also have the same problem with you. The code runs successfully on my Ubuntu20 but failed in Linux. May I know if you have resolved this problem?

uranuscc commented 3 years ago

Hi Williamed,

The root cause in my case is that the gcc version was not compatible with pytorch ver. I have upgraded it from 4.9 to 7+.

Best Regards, Chen Cheng

From: Williamed @.> Sent: Tuesday, May 4, 2021 3:34 AM To: sshaoshuai/PointRCNN @.> Cc: Chen Cheng (IFAP DC PSS SIS SI) @.>; Author @.> Subject: Re: [sshaoshuai/PointRCNN] Segmentation error (#186)

Caution: This e-mail originated outside Infineon Technologies. Do not click on links or open attachments unless you validate it is safehttps://goto.infineon.com/SocialEngineering.

I have tested it on Ubuntu v20.04 with python 3.8, pytorch 1.0. The demo run works fine now. I think it's OS env issue with my RHEL6. cuda version or OS. some incompatibility there. still troubleshooting it...

Hi, I also have the same problem with you. The code runs successfully on my Ubuntu20 but failed in Linux. May I know if you have resolved this problem?

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHubhttps://github.com/sshaoshuai/PointRCNN/issues/186#issuecomment-831483585, or unsubscribehttps://github.com/notifications/unsubscribe-auth/ASLX7HNIJIWPGBEKUH3DUHDTL33DPANCNFSM4VW3VEZA.

QLuanWilliamed commented 3 years ago

Hi， Thanks for your reply. I also resolved it by changing the version of gcc to 5.5.0.

Kind regards, William On 4 May 2021, 02:37 +0100, uranuscc @.***>, wrote:

Hi Williamed,

The root cause in my case is that the gcc version was not compatible with pytorch ver. I have upgraded it from 4.9 to 7+.

Best Regards, Chen Cheng

From: Williamed @.> Sent: Tuesday, May 4, 2021 3:34 AM To: sshaoshuai/PointRCNN @.> Cc: Chen Cheng (IFAP DC PSS SIS SI) @.>; Author @.> Subject: Re: [sshaoshuai/PointRCNN] Segmentation error (#186)

Caution: This e-mail originated outside Infineon Technologies. Do not click on links or open attachments unless you validate it is safehttps://goto.infineon.com/SocialEngineering.

I have tested it on Ubuntu v20.04 with python 3.8, pytorch 1.0. The demo run works fine now. I think it's OS env issue with my RHEL6. cuda version or OS. some incompatibility there. still troubleshooting it...

Hi, I also have the same problem with you. The code runs successfully on my Ubuntu20 but failed in Linux. May I know if you have resolved this problem?

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHubhttps://github.com/sshaoshuai/PointRCNN/issues/186#issuecomment-831483585, or unsubscribehttps://github.com/notifications/unsubscribe-auth/ASLX7HNIJIWPGBEKUH3DUHDTL33DPANCNFSM4VW3VEZA.

— You are receiving this because you commented. Reply to this email directly, view it on GitHubhttps://eur01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fsshaoshuai%2FPointRCNN%2Fissues%2F186%23issuecomment-831636242&data=04%7C01%7C%7Cd4e02440033b41be882908d90e9d2211%7C569df091b01340e386eebd9cb9e25814%7C0%7C0%7C637556890329117695%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C1000&sdata=BFx1%2FL6RCuw6RZRQROnzEYEIgAorwucdyZVsjjCvgSI%3D&reserved=0, or unsubscribehttps://eur01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fnotifications%2Funsubscribe-auth%2FAPBUPY3M7SDMN3FO36OJ7HLTL5FUHANCNFSM4VW3VEZA&data=04%7C01%7C%7Cd4e02440033b41be882908d90e9d2211%7C569df091b01340e386eebd9cb9e25814%7C0%7C0%7C637556890329127691%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C1000&sdata=tIhKRM8mBw8vEGoDSNzNbPE8ne1b3rM9U%2BJ6aY7PkFI%3D&reserved=0.