Closed pythonyyds-byte closed 9 months ago
Hi, you can try decreasing the version of mmcv. The version we used for testing was 1.1.6.
Thank you very much, I have fixed the above error with an updated version. However, when I used BPR to refine the coco validation set, the overall runtime exceeded expectations. Is this normal?
Thank you very much, I have fixed the above error with an updated version. However, when I used BPR to refine the coco validation set, the overall runtime exceeded expectations. Is this normal?
Mr.Right @.***
------------------ 原始邮件 ------------------ 发件人: "chenhang98/BPR" @.>; 发送时间: 2023年9月11日(星期一) 晚上7:04 @.>; @.**@.>; 主题: Re: [chenhang98/BPR] Error in the process of inference (Issue #48)
Hi, you can try decreasing the version of mmcv. The version we used for testing was 1.1.6.
— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: @.***>
Yes, running the entire pipeline of BPR on the COCO validation set might take several hours since there are many instances.
ModuleNotFoundError: No module named 'mmdet.ops' AssertionError: MMCV==1.3.5 is used but incompatible. Please install mmcv>=[1, 1, 4], <=[1, 2, 0]. FileNotFoundError: [Errno 2] No such file or directory: 'mask_rcnn_r50.val.refined.json/refined.pkl'
In the process of reasoning, the above three errors occurred in turn. I suspected that the mmcv-full I used was incompatible with mmdet and mmseg. The following picture is the version of the package I installed. I hope you can explain how to solve this problem.