RandLA-Net

RandLA-Net copied to clipboard

python3.5 utils/data_prepare_s3dis.py

1

(RANDLA-NET) oXXXXX@oXXXXX:~/Documents/RandLA-Net-master$ python3.5 utils/data_prepare_s3dis.py data/S3DIS/Stanford3dDataset_v1.2_Aligned_Version/Area_1/conferenceRoom_1/Annotations data/S3DIS/Stanford3dDataset_v1.2_Aligned_Version/Area_1/conferenceRoom_2/Annotations data/S3DIS/Stanford3dDataset_v1.2_Aligned_Version/Area_1/copyRoom_1/Annotations data/S3DIS/Stanford3dDataset_v1.2_Aligned_Version/Area_1/hallway_1/Annotations data/S3DIS/Stanford3dDataset_v1.2_Aligned_Version/Area_1/hallway_2/Annotations data/S3DIS/Stanford3dDataset_v1.2_Aligned_Version/Area_1/hallway_3/Annotations data/S3DIS/Stanford3dDataset_v1.2_Aligned_Version/Area_1/hallway_4/Annotations data/S3DIS/Stanford3dDataset_v1.2_Aligned_Version/Area_1/hallway_5/Annotations data/S3DIS/Stanford3dDataset_v1.2_Aligned_Version/Area_1/hallway_6/Annotations sys:1: DtypeWarning: Columns (0,5) have mixed types. Specify dtype option on import or set low_memory=False. Traceback (most...

Timaces

while trying to run the data_prepare_S3DIS.py the following error showed: ` Traceback (most recent call last): File "data_prepare_s3dis.py", line 79, in convert_pc2ply(annotation_path, join(original_pc_folder, out_file_name)) File "data_prepare_s3dis.py", line 49, in convert_pc2ply...

youssef962

How to calculate num_per_class?

2

Dear QingyongHu, Congratulations and thanks for your interesting paper RandLA-Net. I am trying to perform it, it is going good. Could you please let me know, in helper_tool.py, page 4-5/6,...

AbdulNur71

Bad results for prediction on full cloud from custom dataset

4

Hi, I'm training the model on my custom dataset which has format similar to SemanticKITTI. I get good results during training on validation set: ``` val: 100%|██████████| 30/30 [00:46

Shershebnev

about experiment on Toronto-3D dataset

1

Hi, @QingyongHu , Could you also upload the code for the experiment on Toronto-3D dataset, as mentioned in your recent PAMI paper? Thanks~

amiltonwong

Compile nearest neighbors in utils folder in windows platform successfully，but the output From test.py in nearest_neighbors does not make sense

6

test code： batch_size = 16 num_points = 81920 K = 16 pc = np.random.rand(batch_size, num_points, 3).astype(np.float32) start = time.time() neigh_idx = nearest_neighbors.knn_batch(pc, pc, K, omp=True) print(neigh_idx) print(time.time() - start) output：...

Asher-1

I have trained and saved the model on Kitti data. While testing it is taking more than 3sec per frame. I am using open3dml for implementation. System has RTX3090 GPU....

bhaskar-anand-iith

运行S3DIS数据时，epoch为0

9

亲爱的胡老师： 您好，我是一名本科生，最近在看您的文章，在运行的时候“sh jobs_6_fold_cv_s3dis.sh”这一条命令时，遇到了 “***EPOCH 0**** Caught a NaN error : 3 2 root error(s) found. (0) Invalid argument: ValueError: attempt to get argmin of an empty sequence Traceback (most recent...

wlwnwafu

Hello,I found a performance issue in the definition of `init_input_pipeline` ,QingyongHu_RandLA-Net/main_S3DIS.py,[self.batch_train_data.map](https://github.com/QingyongHu/RandLA-Net/blob/6b5445f5f279d33d2335e85ed39ca8b68cb1c57e/main_S3DIS.py#L208) was called without **num_parallel_calls**.I think it will increase the efficiency of your program if you add this.Here is [the...

DLPerf

data_prepare_semantic3d.py中存在label的情况下，pickle.dump([proj_idx, labels], f)，labels为什么不和sub_points对应呢

1

wumingbai