video-branch occurs error when followed your steps [No such file or directory: './out/lidar_20Hz/results_nusc.json']

[zhangxiao696]

hello, When I followed your steps, it reported an error, may I ask you how did you successfully execute it?

[flymin]

It seems like ASAP’s problem. Please resort help from them. Similar to #26.

[zhangxiao696]

I have some doubts, ASAP code was last updated two years ago, did you also execute successfully as mentioned above？

[zhangxiao696]

I have some doubts, ASAP code was last updated two years ago, did you also execute successfully as mentioned above？

Their code has been frozen and has not been updated

[flymin]

Yes, we run their codes.

Let me talk more. Our codes do not load their modules but outputs, which means any annotation generation methods should be useful. We just recommend the one we used, which is also adopted by other literatures. Hope it helps.

[zhangxiao696]

Yes, we run their codes.

Let me talk more. Our codes do not load their modules but outputs, which means any annotation generation methods should be useful. We just recommend the one we used, which is also adopted by other literatures. Hope it helps.

okey, I get, thanks you a lot

[potentialming]

Yes, we run their codes. Let me talk more. Our codes do not load their modules but outputs, which means any annotation generation methods should be useful. We just recommend the one we used, which is also adopted by other literatures. Hope it helps.

okey, I get, thanks you a lot

Did you finally solve the problem? Did you successfully complete the training? I successfully generated interp_12Hz_trainval by commenting out the relevant code of nusc_annotation_generator.py, but when I ran bash scripts/prepare_dataset.sh, an error occurred KeyError: 'a5f12393fdf042a4a1c25cbe29f3f4ea'. Is it because the environment of the main branch was used when using the video branch? Hope to get your help.

[LeningWang]

Yes, we run their codes. Let me talk more. Our codes do not load their modules but outputs, which means any annotation generation methods should be useful. We just recommend the one we used, which is also adopted by other literatures. Hope it helps.

okey, I get, thanks you a lot

Did you finally solve the problem? Did you successfully complete the training? I successfully generated interp_12Hz_trainval by commenting out the relevant code of nusc_annotation_generator.py, but when I ran bash scripts/prepare_dataset.sh, an error occurred KeyError: 'a5f12393fdf042a4a1c25cbe29f3f4ea'. Is it because the environment of the main branch was used when using the video branch? Hope to get your help.

I have same problem

[flymin]

For you guys still having this issue. We will release the processed data annotations tmr, please wait for the updates in w-coda2024 track2.

[flymin]

@potentialming Your problem is clearly environment issue as shown in the last line. Please install the mmdet3d located in the video branch.

@LeningWang See above.

BTW, this problem is off-topic for this issue. Any follow-ups should go to other issues or a new one. Thanks.

[dyy0205]

Yes, we run their codes. Let me talk more. Our codes do not load their modules but outputs, which means any annotation generation methods should be useful. We just recommend the one we used, which is also adopted by other literatures. Hope it helps.

okey, I get, thanks you a lot

Did you finally solve the problem? Did you successfully complete the training? I successfully generated interp_12Hz_trainval by commenting out the relevant code of nusc_annotation_generator.py, but when I ran bash scripts/prepare_dataset.sh, an error occurred KeyError: 'a5f12393fdf042a4a1c25cbe29f3f4ea'. Is it because the environment of the main branch was used when using the video branch? Hope to get your help. Have you solved the KeyError problem?