tsunghan-wu / randla-net-pytorch Goto Github PK

Hello, thanks for your amazing work.

After the model had trained. I using test_SemanticKITTI.py for inference. However, I found the self.test_dataset.min_possibility is not updating during the test time. Could you please give me some suggestions?

when i use this code: python3 train_SemanticKITTI.py --checkpoint_path pretrain_model/checkpoint.tar,
This error will happen: confusion_matrix() takes 2 positional arguments but 3 were given.
could you tell me how to modify this code?

Which path should I set about dst_path ? Can you give me an example ？
And I cant find sequences_0.06 after I run data_prepare_semantickittii.py .
Can you help me ?
Thank you very much .

I used 2 GPUs on a single machine to train model, and it correctly worked. However, when using more 2 GPUs for training, the program stucked in network's forward propagation step, where the utilization rate of GPUs stays 100%

@tsunghan-mama Hi! Thank you for your excellent work! I am trying to infer the sequence 08 by using the pre-trained model. However, an error occurred: ModuleNotFoundError: No module named 'sklearn.metrics._dist_metrics'.
I search the documents of pickle in [https://scikit-learn.org/stable/search.html?], and no version has the module"sklearn.metrics._dist_metrics". Could you please provide some suggestions to solve this problem?
Looking forward to your reply!
Thank you so much!

More Details about the error:
Traceback (most recent call last):
File "test_SemanticKITTI.py", line 180, in
main()
File "test_SemanticKITTI.py", line 176, in main
tester.test()
File "test_SemanticKITTI.py", line 99, in test
self.rolling_predict()
File "test_SemanticKITTI.py", line 110, in rolling_predict
batch_data, input_inds, cloud_inds, min_possibility = next(iter_loader)
File "//python3.7/site-packages/torch/utils/data/dataloader.py", line 345, in next
data = self._next_data()
File "//python3.7/site-packages/torch/utils/data/dataloader.py", line 385, in _next_data
data = self._dataset_fetcher.fetch(index) # may raise StopIteration
File "//python3.7/site-packages/torch/utils/data/_utils/fetch.py", line 34, in fetch
data = next(self.dataset_iter)
File "//RandLA-Net-pytorch-main/dataset/semkitti_testset.py", line 52, in spatially_regular_gen
pc, tree, labels = self.get_data(pc_path)
File "/***/RandLA-Net-pytorch-main/dataset/semkitti_testset.py", line 69, in get_data
search_tree = pickle.load(f)
ModuleNotFoundError: No module named 'sklearn.metrics._dist_metrics'

Hi, I would like to know how much memory you need for testing SemanticKITTI. When setting batch=1, I need almost 32G of memory (not GPU memory). Is this normal? Or is there any way to reduce that demand？

code errors in utils.semkitti_vis.laserscan.py

Hi, it would crash when I run train_SemanticKITTI.py. I change to cpu mode debug and infomation like following:

How to solve this? @tsunghan-mama @dream-toy @huixiancheng

run data_prepare_semantickitti.py ,and there's an error: ModuleNotFoundError: No module named 'utils.nearest_neighbors.lib'
After check, there is no such a file.

你好，训练时有53.9的iou，但是测试完用evaluate_SemanticKITTI.py 估计时却是如下情况。测试显示的iou和最后的差距大吗

`validation set:
Acc avg 0.052
IoU avg 0.003
IoU class 1 [car] = 0.052
IoU class 2 [bicycle] = 0.000
IoU class 3 [motorcycle] = 0.000
IoU class 4 [truck] = 0.000
IoU class 5 [other-vehicle] = 0.000
IoU class 6 [person] = 0.001
IoU class 7 [bicyclist] = 0.000
IoU class 8 [motorcyclist] = 0.000
IoU class 9 [road] = 0.000
IoU class 10 [parking] = 0.002
IoU class 11 [sidewalk] = 0.000
IoU class 12 [other-ground] = 0.000
IoU class 13 [building] = 0.000
IoU class 14 [fence] = 0.000
IoU class 15 [vegetation] = 0.000
IoU class 16 [trunk] = 0.000
IoU class 17 [terrain] = 0.000
IoU class 18 [pole] = 0.000
IoU class 19 [traffic-sign] = 0.000

0.052,0.000,0.000,0.000,0.000,0.001,0.000,0.000,0.000,0.002,0.000,0.000,0.000,0.000,0.000,0.000,0.000,0.000,0.000,0.003,0.052`
不知道用中文简体字提问能看懂吗。

I have tried using Python 3.10, 3.8, and 3.7, but the compilation with the bash script 'compile_op.sh' failed for all of them. Only version 3.5 was successful. Does this project specifically require Python version 3.5 as well?

How to conduct online test

I use the pre-trained model to test the sequence 08, and i get 52.9% mean IoU but i find the visualization seems work badly

pickle.UnpicklingError: A load persistent id instruction was encountered,
but no persistent_load function was specified.

when i use pretrain_models torch.load error. So I want to know whats the torch version and cuda version.Thank you very much

I got a bug when I training

(randlanet) wx@dl-group-workstation:/media/wx/HDD/DQ/RandLA-Net-pytorch-main$ python train_SemanticKITTI.py
100%|█████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 19130/19130 [00:40<00:00, 473.39it/s]
0%| | 0/3826 [00:03<?, ?it/s]
Traceback (most recent call last):
File "train_SemanticKITTI.py", line 191, in
main()
File "train_SemanticKITTI.py", line 187, in main
trainer.train()
File "train_SemanticKITTI.py", line 131, in train
self.train_one_epoch()
File "train_SemanticKITTI.py", line 120, in train_one_epoch
loss, end_points = compute_loss(end_points, self.train_dataset, self.criterion)
File "/media/wx/HDD/DQ/RandLA-Net-pytorch-main/network/loss_func.py", line 28, in compute_loss
loss = criterion(valid_logits, valid_labels).mean()
File "/home/wx/anaconda3/envs/randlanet/lib/python3.6/site-packages/torch/nn/modules/module.py", line 1102, in _call_impl
return forward_call(*input, **kwargs)
File "/home/wx/anaconda3/envs/randlanet/lib/python3.6/site-packages/torch/nn/modules/loss.py", line 1152, in forward
label_smoothing=self.label_smoothing)
File "/home/wx/anaconda3/envs/randlanet/lib/python3.6/site-packages/torch/nn/functional.py", line 2846, in cross_entropy
return torch._C._nn.cross_entropy_loss(input, target, weight, _Reduction.get_enum(reduction), ignore_index, label_smoothing)
RuntimeError: weight tensor should be defined either for all 19 classes or no classes but got weight tensor of shape: [1, 19]

could you please help me？

Hello, I want to use the point cloud I collected to test the pre-trained model on semanticKITTI. I just convert the ply file to .npy and put it in velodyne, but it will show that the array is out of bounds. I need to run dataset prepare to generate a new KDTree and proj files?

As in follow, the left two is the number of points infer by the test_SemanticKITTI.py. The right is the number of points from orgin SemanticKITTI .

Maybe missing the proj.pkl part in val or test part?
https://github.com/QingyongHu/RandLA-Net/blob/6b5445f5f279d33d2335e85ed39ca8b68cb1c57e/tester_SemanticKITTI.py#L109-L114

I have trained 30 epoches. The best accuracy and iou is only 33.75% and 23.16%, respectively.
I have used 1800 samples in Nuscense dataset. Each sample contains 34000 points.
The learning rate is set to 0.001. CosineAnnealingLR strategy is token..
I wonder if anything wrong with me. Thank you very much.

hi, tanks you for your great work! I run the test, but the min_possibility always 0, have you this problem, or why this happen? thank for your reply in your free time.

when I run python train_SemanticKITTI.py ,
an error happen:
100%|███████████████████████████████████████| 4541/4541 [04:12<00:00, 17.99it/s]
100%|█████████████████████████████████████████| 909/909 [18:11<00:00, 1.20s/it]
0%| | 0/136 [00:08<?, ?it/s]
Traceback (most recent call last):
File "train_SemanticKITTI.py", line 197, in
main()
File "train_SemanticKITTI.py", line 193, in main
trainer.train()
File "train_SemanticKITTI.py", line 138, in train
mean_iou = self.validate()
File "train_SemanticKITTI.py", line 167, in validate
iou_calc.add_data(end_points)
File "/home/tukrin/ZYD_3D/RandLA-Net-pytorch-main/utils/metric.py", line 29, in add_data
conf_matrix = confusion_matrix(labels_valid, pred_valid, np.arange(0, self.cfg.num_classes, 1))
TypeError: confusion_matrix() takes 2 positional arguments but 3 were given

I don"t know how to solve this problem,can you help me? thanks!!!

        if mean_iou > self.highest_val_iou:
            self.hightest_val_iou = mean_iou

code there wrong

I try to train it by the way of ReadMe.md,But there are some problems:
File "/home/xavier/RidarSS/RandLA-Net-pytorch-master/helper_tool.py", line 171, in knn_search neighbor_idx = nearest_neighbors._knn_batch(support_pts, query_pts, k, omp=True) AttributeError: module 'nearest_neighbors' has no attribute 'knn_batch'
any solution?plz

Thank you for the realization, great job. I was wondering if you submitted the seq11-21 prediction results to test codalab?
As far as I know, there is a large difference between the valid and test of the Semantickitti dataset.

Firstly, thank you for your work, an up-to-date pytorch implementation of RandLA is really nice to have.
This is not a bug report but rather a series of questions I had when I started implementing RandLA before I found this repository.

• The loss weights in Pytorch CrossEntropy are normalized so that they sum to one, I don't think the original code has this normalization, but removing it would generate huge loss values.
• Why use ConvLayers, since they all have 1x1 kernels, Linear layers would work just as well?
• regarding https://github.com/qiqihaer/RandLA-Net-pytorch/blob/a255652b65d8378682479c23299704783a8fe4d9/RandLANet.py#L58 it matches the tensorflow code but not the fig. 7 of the paper where the shortcut originates from before the first res block, right?

After i've run data_preprare_semantickitti.py , I want to see how the pointcloud looks like.
So i run command -- python .\visualize_SemanticKITTI.py -d .\dataset\semantickitti\sequences_0.06\ -s 03
But there is an error
RuntimeError: Filename extension is not valid label file.

I noticed that it requires label file needed to has a .label filename extension.
But after data_prepare, my pointcloud&label file have the extension of .npy:
self.scan_names[self.offset] is^
.\dataset\semantickitti\sequences_0.06\03\velodyne\000000.npy^
self.label_names[self.offset] is^
.\dataset\semantickitti\sequences_0.06\03\labels\000000.npy^

So i try to visualize raw data, i run python .\visualize_SemanticKITTI.py -d .\dataset\semantickitti\sequences\ -s 03
But there is also an error:
RuntimeError: Filename extension is not valid scan file.
It requirs .npy, but raw kitii dataset has pointcloud on .bin while label on .label

So how can i run visualization?

RuntimeError: weight tensor should be defined either for all 19 classes or no classes but got weight tensor of shape: [1, 19]
how to solve