Mismatch between predicted points and GT points during testing → IndexError in unified_metric.py

[giserhaili]

Hi, thanks for your great work.
When reproducing the paper using the provided config, I encountered an error during testing caused by a mismatch between predicted points and GT points.

Error Message：
Traceback (most recent call last):
File "/workspace/tools/test.py", line 176, in
main()
File "/workspace/tools/test.py", line 172, in main
runner.test()
File "/opt/conda/lib/python3.10/site-packages/mmengine/runner/runner.py", line 1767, in test
metrics = self.test_loop.run() # type: ignore
File "/opt/conda/lib/python3.10/site-packages/mmengine/runner/loops.py", line 442, in run
metrics = self.evaluator.evaluate(len(self.dataloader.dataset))
File "/opt/conda/lib/python3.10/site-packages/mmengine/evaluator/evaluator.py", line 79, in evaluate
_results = metric.evaluate(size)
File "/opt/conda/lib/python3.10/site-packages/mmengine/evaluator/metric.py", line 133, in evaluate
_metrics = self.compute_metrics(results) # type: ignore
File "/workspace/oneformer3d/unified_metric.py", line 96, in compute_metrics
gt_l, pred_l = int(sem_gt_i[j]), int(sem_pre_i[j])
IndexError: index 20736 is out of bounds for axis 0 with size 20736

[anloganathan]

Hi, I am also facing the same issue. But the computed .ply files would have already got saved locally, so we can do the metric computation part explicitly according to this comment: #13 (comment)

[anloganathan]

And after computing the metrics explictly, I am facing same issue as this: #15 (comment)

But after switching to the commit before november, it worked.

[giserhaili]

And after computing the metrics explictly, I am facing same issue as this: #15 (comment)在明确计算了这些指标后，我遇到了同样的问题： #15 (comment)

But after switching to the commit before november, it worked.但在11月之前切换到提交后，它就正常了。

Hello, thank you very much for your reply, it did work. Next, should I change the code in the green area to the code in the red area?

[anloganathan]

Hi, if you run the below command inside your output folder which you have used in your explicit final_eval run, you might just get -1 and nothing else, which means instance segmentation prediction isn't correct.

sed '1,/end_header/d' *.ply | awk '{print $5}' | sort -n | uniq

So, I checked out the entire repo to the commit id: 83f9bbc, and then did the whole inference again, and then ran the final_eval explictly again. It worked.

[anloganathan]

Hi, I have tried with the above mentioned old commit. And, I am able to successfully visualize the predictions in cloud compare.

[bonuli123]

你好，我已经尝试过上面提到的旧提交。而且，我能够成功地在云端比较中可视化预测。

Hello, I want to ask you, when did you use the warehouse before? Have you done any pre-training using the data given by the author? There is nothing wrong with my inference part, and there are correct results, but my pre-training has always been problematic.