IndexError: boolean index did not match indexed array along dimension 0; dimension is 1199 but corresponding boolean dimension is 127533

[Ckeai]

Hello, I am very interested in your work. While replicating your code, I encountered an issue, as shown in the attached image. During training, the dimensions of valid_idxs do not match with the dimensions of the mean label mu_label. The mu_label is generated in the pseudo-label generation part of the code. This issue has been troubling me for quite some time, and I would appreciate your guidance on this matter.

Traceback (most recent call last):
File "tools/train.py", line 351, in
main()
File "tools/train.py", line 342, in main
train(epoch, model, optimizer, scheduler, scaler, train_loader, cfg, logger, writer)
File "tools/train.py", line 64, in train
for i, batch in enumerate(train_loader, start=1):
File "/WANGboyue/wbw/anaconda3/envs/gapro_backup/lib/python3.8/site-packages/torch/utils/data/dataloader.py", line 681, in next
data = self._next_data()
File "/WANGboyue/wbw/anaconda3/envs/gapro_backup/lib/python3.8/site-packages/torch/utils/data/dataloader.py", line 1376, in _next_data
return self._process_data(data)
File "/WANGboyue/wbw/anaconda3/envs/gapro_backup/lib/python3.8/site-packages/torch/utils/data/dataloader.py", line 1402, in _process_data
data.reraise()
File "/WANGboyue/wbw/anaconda3/envs/gapro_backup/lib/python3.8/site-packages/torch/_utils.py", line 461, in reraise
raise exception
IndexError: Caught IndexError in DataLoader worker process 0.
Original Traceback (most recent call last):
File "/WANGboyue/wbw/anaconda3/envs/gapro_backup/lib/python3.8/site-packages/torch/utils/data/_utils/worker.py", line 302, in _worker_loop
data = fetcher.fetch(index)
File "/WANGboyue/wbw/anaconda3/envs/gapro_backup/lib/python3.8/site-packages/torch/utils/data/_utils/fetch.py", line 49, in fetch
data = [self.dataset[idx] for idx in possibly_batched_index]
File "/WANGboyue/wbw/anaconda3/envs/gapro_backup/lib/python3.8/site-packages/torch/utils/data/_utils/fetch.py", line 49, in
data = [self.dataset[idx] for idx in possibly_batched_index]
File "/WANGboyue/wbw/GaPro-master/ISBNet/isbnet/data/custom.py", line 174, in getitem
self.transform_train(xyz, rgb, semantic_label, instance_label, prob_label, mu_label, var_label, spp)
File "/WANGboyue/wbw/GaPro-master/ISBNet/isbnet/data/custom.py", line 148, in transform_train
mu_label = mu_label[valid_idxs]
IndexError: boolean index did not match indexed array along dimension 0; dimension is 1199 but corresponding boolean dimension is 127533

[ngoductuanlhp]

Could you double-check the initial shape of the semantic_label, instance_label, prob_label, mu_label when loading at here? If you process the pseudo-label correctly, all these tensors have the similar shape of (N) where N is the total number of points.

[Ckeai]

I followed the readme, but the variance and mean of the processed pseudo-labels are indeed incorrect in dimension.

[Stay-Naive]

I came across the same problem. Do you have any idea now to solve this problem? @ngoductuanlhp @Ckeai
By the way, when I check the code in gen_ps.py, the wall_box confused me. In gapro/scannet_planes.py/get_wall_boxes, this part of data seems to come from dataset/scannetv2/scannet_planes/. However, I downloaded and prepared the scannet dataset following the document, and I did not have this part of data. Also, I did not see any instruction about this.

[zyjhxd]

我遇到了同样的问题。您现在有什么想法可以解决这个问题吗？@ngoductuanlhp @Ckeai 顺便说一句，当我检查中的代码时gen_ps.py，让wall_box我感到困惑。在中gapro/scannet_planes.py/get_wall_boxes，这部分数据似乎来自dataset/scannetv2/scannet_planes/。但是，我按照文档下载并准备了 scannet 数据集，但我没有这部分数据。而且，我没有请参阅有关此的任何说明。

您好请问解决了吗

[zyjhxd]

我按照自述文件进行操作，但处理后的伪标签的方差和均值在维度上确实不正确。

您好请问解决了吗