laiqingsi / clocs_lqs Goto Github PK

Hallo Laiqingsi,

Thanks for your great implementation. Could you provide the 3D detection results from SECOND? The link you post in readme is not reachable.

您好，我目前已经完成了OpenPCDet的运行（在其他文件夹中），从上交云盘中下载了您提供的2D与3D目标检测结果，可是CLOCs工程中没有setup.py，使得它不能被安装并运行，我应该如何处理呢？真心需要您的帮助，可以麻烦添加一下联系方式吗？我的邮箱是[email protected]

Hello, I have finished running OpenPCDet (in other folders) and downloaded your 2D and 3D target detection results from the Cloud, but there is no setup.py in the CLOCs project, which makes it impossible to install and run. What should I do? I really need your help. My E-mail is [email protected], would you like to add some contact information?

非常感谢你更新了readme，我也已经再次尝试运行你的出色工作。但是还是有一些问题导致无法运行：
1、在CLOCs_LQS/generate_data.py中，是否需要做如下修改，否则报错
if name == "main":
args, cfg = parse_args()
_2d_path = args.d2path
_3d_path = args.d3path
d2method=args.d2method
d3method=args.d3method
input_data = args.inputpath
train_ind_path = args.train_indexpath
val_ind_path = args.val_indexpath

train_dataset = clocs_data(_2d_path, _3d_path,train_ind_path, input_data, d2method , d3method)
train_dataset.generate_input()

2、是否需要用安装后的OpenPCDet工程中替换如下几个文件，否则报错：
/CLOCs_LQS/pcdet/ops/roipoint_pool3d/roipoint_pool3d_cuda.cpython-37m-x86_64-linux-gnu.so……
/CLOCs_LQS/tool/pcdet/ops/roipoint_pool3d/roipoint_pool3d_cuda.cpython-37m-x86_64-linux-gnu.so……

3、我完成以上修改后，运行generate_data.py出现了以下的报错，我要如何继续修改呢：
home/wenhao/anaconda3/envs/openpcdet/lib/python3.7/site-packages/numba/cuda/decorators.py:110: NumbaDeprecationWarning: Eager compilation of device functions is deprecated (this occurs when a signature is provided)
warn(NumbaDeprecationWarning(msg))
Traceback (most recent call last):
File "generate_data.py", line 62, in
train_dataset = clocs_data(_2d_path, _3d_path,train_ind_path, input_data, d2method , d3method)
File "/home/wenhao/桌面/LQS/CLOCs_LQS/tool/dataset.py", line 28, in init
IndexError: only integers, slices (:), ellipsis (...), numpy.newaxis (None) and integer or boolean arrays are valid indices

……………………………………………………………………………………………………………………………………………………………………………………………………

Thank you very much for updating readme, and I've tried to run your great work again. However, there are still some problems that make it impossible to run:
1、In CLOCs_LQS/generate_data.py , should I need to make the following modifications, otherwise an error will be reported?
if name == "main":
args, cfg = parse_args()
_2d_path = args.d2path
_3d_path = args.d3path
d2method=args.d2method
d3method=args.d3method
input_data = args.inputpath
train_ind_path = args.train_indexpath
val_ind_path = args.val_indexpath

train_dataset = clocs_data(_2d_path, _3d_path,train_ind_path, input_data, d2method , d3method)
train_dataset.generate_input()

2、Should I need to replace the following files with the installed OpenPCDet project, otherwise an error will occur:
/CLOCs_LQS/pcdet/ops/roipoint_pool3d/roipoint_pool3d_cuda.cpython-37m-x86_64-linux-gnu.so……
/CLOCs_LQS/tool/pcdet/ops/roipoint_pool3d/roipoint_pool3d_cuda.cpython-37m-x86_64-linux-gnu.so……

3、After I have made the above changes, run generate_data.py The following error occurred. How do I proceed with the modification?
home/wenhao/anaconda3/envs/openpcdet/lib/python3.7/site-packages/numba/cuda/decorators.py:110: NumbaDeprecationWarning: Eager compilation of device functions is deprecated (this occurs when a signature is provided)
warn(NumbaDeprecationWarning(msg))
Traceback (most recent call last):
File "generate_data.py", line 62, in
train_dataset = clocs_data(_2d_path, _3d_path,train_ind_path, input_data, d2method , d3method)
File "/home/wenhao/桌面/LQS/CLOCs_LQS/tool/dataset.py", line 28, in init
IndexError: only integers, slices (:), ellipsis (...), numpy.newaxis (None) and integer or boolean arrays are valid indices

I trained a second_car network , it works well.

when I use the second-predict.yaml to get no NMS resluts, I got the scores like this:

So,how to get no NMS results in pcdet? Could you share the skills ,please?

Thanks for your brilliant work and I wonder whether it can support fusion for pedestrian and cyclists on KITTI?

When I meet fellow error,
2022-03-14 15:50:24,384 INFO Generate output done Traceback (most recent call last): File "train.py", line 191, in eval(fusion_layer, val_data, logf, log_path, epoch, cfg, eval_set, logger) File "train.py", line 125, in eval eval_metric=cfg.MODEL.POST_PROCESSING.EVAL_METRIC File "/home/wenhao/桌面/LQS/CLOCs_LQS/pcdet/datasets/kitti/kitti_dataset.py", line 323, in evaluation if 'annos' not in self.kitti_infos[0].keys(): IndexError: list index out of range

And I find that coda in /CLOCs_LQS/pcdet/datasets/kitti/kitti_dataset.py,line 323,
def evaluation(self, det_annos, class_names, **kwargs): if 'annos' not in self.kitti_infos[0].keys(): return None, {}
self.kitti_infos is empyt [ ] , then self.kitti_infos[0] out of range.Cloud you please tell me how can I solve this error?

Hi @Laiqingsi, thanks for your great work. It helps me a lot!

However, I found that the AP improvement by fusion is very low (even degradation in some cases) compared to original CLOCs implementation. For example, when I train the fusion model with my own 2d detection results from Retinanet in Detectron2, it doesn't gain anything from fusion. Oppositely, I can get a 3%+ improvement in the original implementation.

Results based on your code:

LiDAR-only:
Car [email protected], 0.70, 0.70:
bbox AP:90.8318, 90.0370, 89.2701
bev  AP:90.0336, 88.1284, 86.8221
3d   AP:88.6435, 78.5305, 77.4470
aos  AP:90.82, 89.93, 89.11
Car [email protected], 0.70, 0.70:
bbox AP:96.1435, 94.8264, 92.1872
bev  AP:92.8031, 90.3776, 88.0352
3d   AP:90.6124, 81.9148, 79.0518
aos  AP:96.13, 94.69, 92.00

CLOCs:
Car [email protected], 0.70, 0.70:
bbox AP:94.2373, 90.3053, 89.6129
bev  AP:89.8845, 88.1735, 87.0562
3d   AP:88.3369, 78.4858, 77.3919
aos  AP:94.20, 90.22, 89.47
Car [email protected], 0.70, 0.70:
bbox AP:97.5323, 94.8859, 93.8404
bev  AP:93.5442, 90.4057, 87.8485
3d   AP:90.2475, 81.7896, 78.7422
aos  AP:97.50, 94.77, 93.64

It's quite strange as I can't find an obvious difference between your code and the original implementation. I wonder that have you encountered such an issue when evaluating different 2d predictions? What may be the possible reason?

Notes: I've also tried to increase the 3D prediction number to 70400 for the second model, but got a similar observation.

Hi,thank you for your work.
But some error occurs,I hope you can give me some suggestions.
When I train the model.the 3D result is always
bbox AP:0.0000, 0.0000, 0.0000
bev AP:0.0000, 0.0000, 0.0000
3d AP:0.0000, 0.0000, 0.0000
Car [email protected], 0.70, 0.70:
bbox AP:0.0000, 0.0000, 0.0000
bev AP:0.0000, 0.0000, 0.0000
3d AP:0.0000, 0.0000, 0.0000
Car [email protected], 0.50, 0.50:
bbox AP:0.0000, 0.0000, 0.0000
bev AP:0.0000, 0.0000, 0.0000
3d AP:0.0000, 0.0000, 0.0000
Car [email protected], 0.50, 0.50:
bbox AP:0.0000, 0.0000, 0.0000
bev AP:0.0000, 0.0000, 0.0000
3d AP:0.0000, 0.0000, 0.0000

您好，非常感谢您的工作，我也希望能在OpenPCDet工程下完成对CLOCs的复现，请问能不能更具体的描述一下您的项目代码要如何运行？数据要如何放置？

Hello, thank you very much for your work, and I would like to complete the CLOCs replication under the OpenPCDet project. Can you describe more specifically how can I run your project code works? How to place the KITTI data?

Hello, i tried this technique to compute clocs and the results are really bad, Can someone help me with this?

2023-02-06 14:49:50,138 INFO Car [email protected], 0.70, 0.70:
bbox AP:4.2648, 4.2198, 3.2463
bev AP:4.1331, 4.0580, 3.1242
3d AP:4.0527, 3.8060, 2.9328
aos AP:4.26, 4.18, 3.21
Car [email protected], 0.70, 0.70:
bbox AP:4.4958, 4.3517, 3.2733
bev AP:4.1676, 4.1848, 3.1503
3d AP:4.0865, 3.6632, 2.6884
aos AP:4.49, 4.31, 3.24
Car [email protected], 0.50, 0.50:
bbox AP:4.2648, 4.2198, 3.2463
bev AP:4.2682, 4.2643, 3.2814
3d AP:4.2680, 4.2537, 3.2734
aos AP:4.26, 4.18, 3.21
Car [email protected], 0.50, 0.50:
bbox AP:4.4958, 4.3517, 3.2733
bev AP:4.4994, 4.3975, 3.3087
3d AP:4.4992, 4.3866, 3.3007
aos AP:4.49, 4.31, 3.24

Thanks for your great work. I have a question, do both 2D detection and 3D detection require the results before NMS, can you elaborate on how to obtain the 2D and 3D detection results before NMS, I would be very grateful.

Hello,

I was curious of how you are generating .pt format files from openpcdet as i'm getting the output as .pkl file.