huangkuns / wireframe Goto Github PK

I've been trying to recreate my own pickle files with the same format as yours. I was wondering if you could point me to some documentation that explains the variables: pointlines, points,
pointlines_index, lines, img, junction, theta, theta_lineIndex, and imgname?

I tried to use the wireframe.py in a single image, but there are not any junction or line result in the result image. I think the cause of this problem is that there is no .npy file corresponding to the image in the line folder, but I don't know how to solve this problem. Has anyone encountered a similar situation, thanks in advance！

Hello, the data download link seems to be invalid.

Dear @huangkuns ,
what's the license of the dataset? Can it be used for commercial purposes?

Cheers

请问可以重新给一个数据下载连接吗？
百度网盘实效了

#Val images: 462
=> Checking checkpoints
=> Loading the latest checkpoint ../result/linepx/latest.pth.tar
=> Creating model from file: models/stackedHGB.py
=> Resuming model state from ../result/linepx/model_200.pth.tar
=> Resuming optimizer state from ../result/linepx/optimState_200.pth.tar
=> Resuming criterion from ../result/linepx/criterion_200.pth.tar
Traceback (most recent call last):
File "main.py", line 85, in
main()
File "main.py", line 52, in main
loss = trainer.test(valLoader, 0)
File "/home/xh/wireframe/linepx/models/stackedHGB-train.py", line 85, in test
for i, (inputData, line, imgids) in enumerate(valLoader):
File "/usr/lib/anaconda3/lib/python3.6/site-packages/torch/utils/data/dataloader.py", line 345, in next
data = self._next_data()
File "/usr/lib/anaconda3/lib/python3.6/site-packages/torch/utils/data/dataloader.py", line 385, in _next_data
data = self._dataset_fetcher.fetch(index) # may raise StopIteration
File "/usr/lib/anaconda3/lib/python3.6/site-packages/torch/utils/data/_utils/fetch.py", line 47, in fetch
return self.collate_fn(data)
File "/usr/lib/anaconda3/lib/python3.6/site-packages/torch/utils/data/_utils/collate.py", line 79, in default_collate
return [default_collate(samples) for samples in transposed]
File "/usr/lib/anaconda3/lib/python3.6/site-packages/torch/utils/data/_utils/collate.py", line 79, in
return [default_collate(samples) for samples in transposed]
File "/usr/lib/anaconda3/lib/python3.6/site-packages/torch/utils/data/_utils/collate.py", line 55, in default_collate
return torch.stack(batch, 0, out=out)
RuntimeError: invalid argument 0: Sizes of tensors must match except in dimension 0. Got 335 and 500 in dimension 2 at /pytorch/aten/src/TH/generic/THTensor.cpp:612

Hello,
Thanks for your excellent work and dataset!

I just want to annotate new data as wireframe format and, I have already read the answer about data.

In labelme program, I can save the points of line, but I cannot revise saving data same as your data structure.

Could you explain more details please? It is my first time, I don't know well.

At last, I want to know how to calculate junction and theta.

Thanks a lot.

how to convert Junctions model to ncnn?

hi，Excuse me, what's the problem?
when I running "python main.py --create_dataset --exp 1 --json":

File "F:\机器学习\源代码\wireframe-master\wireframe-master\junc\datasets\utils.py", line 333, in load_data_and_save
bar = pb.ProgressBar(widgets=[ '[ ', pb.Bar(), ' ][ ', pb.Timer(), ' ]'], max_value=pb.UnknownLength)
TypeError: init() got an unexpected keyword argument 'max_value'

It seems that the link is unusable

Hi,
I'd like to have a quick verification with my own dataset based on your "wireframe".
1.
However, I cant find the any pre-trained models(*.pkl or *.pth files). Can you release them?
2.
During the inference stage, will the output image size of arbitrary input images be unified to 320.*320(default imgdim in opts.py )?
3.
Do you have plans to support the panorama images?
4.
If I have to train from scratch, I'd like to know how long it will take to train on single GPU of P4 processor ?

Best Regards,
Jin

When trying to execute:
python3 main.py --create_dataset --exp 1 --json

I got the following error:

#train images: 5000
src_dir: /home/wireframe/junc/../data/pointlines save_dir: /home/wireframe/junc/../data/junc/processed
== loading raw data ==
[ |######################################################################| ][ Elapsed Time: 0:00:05 ]
== 5000 raw images data loaded ==
joblib.externals.loky.process_executor._RemoteTraceback:                 | ][ Elapsed Time: 0:00:01 ]
"""
Traceback (most recent call last):
  File "/home/.virtualenvs/wireframe/lib/python3.6/site-packages/joblib/externals/loky/process_executor.py", line 418, in _process_worker
    r = call_item()
  File "/home/.virtualenvs/wireframe/lib/python3.6/site-packages/joblib/externals/loky/process_executor.py", line 272, in __call__
    return self.fn(*self.args, **self.kwargs)
  File "/home/.virtualenvs/wireframe/lib/python3.6/site-packages/joblib/_parallel_backends.py", line 608, in __call__
    return self.func(*args, **kwargs)
  File "/home/.virtualenvs/wireframe/lib/python3.6/site-packages/joblib/parallel.py", line 256, in __call__
    for func, args, kwargs in self.items]
  File "/home/.virtualenvs/wireframe/lib/python3.6/site-packages/joblib/parallel.py", line 256, in <listcomp>
    for func, args, kwargs in self.items]
  File "/home/wireframe/junc/datasets/utils.py", line 315, in save_ann
    os.mkdir(dir_to_save)
FileNotFoundError: [Errno 2] No such file or directory: '/home/wireframe/junc/../data/junc/processed/480_60_15'
"""

The above exception was the direct cause of the following exception:

Traceback (most recent call last):
  File "main.py", line 144, in <module>
    main()
  File "main.py", line 105, in main
    if check_dataset(opt, 'train'):
  File "main.py", line 94, in check_dataset
    create_dataset(H, split, use_mp=True); 
  File "/home/wireframe/junc/datasets/utils.py", line 417, in create_dataset
    load_data_and_save(H, src_dir=src_dir, save_dir=save_dir, filenames=filelst, use_mp = use_mp)
  File "/home/wireframe/junc/datasets/utils.py", line 349, in load_data_and_save
    for d in bar(anno))  # max_len default to 1.
  File "/home/.virtualenvs/wireframe/lib/python3.6/site-packages/joblib/parallel.py", line 1017, in __call__
    self.retrieve()
  File "/home/.virtualenvs/wireframe/lib/python3.6/site-packages/joblib/parallel.py", line 909, in retrieve
    self._output.extend(job.get(timeout=self.timeout))
  File "/home/.virtualenvs/wireframe/lib/python3.6/site-packages/joblib/_parallel_backends.py", line 562, in wrap_future_result
    return future.result(timeout=timeout)
  File "/home/anaconda3/lib/python3.6/concurrent/futures/_base.py", line 432, in result
    return self.__get_result()
  File "/home/anaconda3/lib/python3.6/concurrent/futures/_base.py", line 384, in __get_result
    raise self._exception
FileNotFoundError: [Errno 2] No such file or directory: '/home/wireframe/junc/../data/junc/processed/480_60_15'

Any idea how to fix the dir path issue? Thank you

I am trying to run the evaluation script, but it requires a folder for ground truth labels of the junctions that does not seem to be present in any of the data that I have downloaded. Is this just in some location that I do not expect, or did this get left out from the datasets somehow?

Hello,
Thanks for your excellent work!

But the one drive link you provided is invalid, so it is impossible for me to download the data. Could you please offer a new link? Maybe on google drive or any other platform.

Thank you very much!

Weipeng WEI

I want to ask how this .pkl dataset is generated, and which annotation tool should be used to annotate data

Hello, may I ask，where can download the source code of ACJ and MJ

Dataset links(baidu & onedrive) are unavailable now, could you update？
Thanks in advance.
Solved:
one drive link is available on Windows(Linux 1604 unavailable)

Following the instructions on the page, when I get to the section to run wireframe.py I get the error: "No such file or directory: 'result/linepx/1/00031546_line.npy'"

As far as I can tell from the code none of the files in the linepx folder reference the result directory, nor do these files seem to be output to any other directory.

Hey, thanks for your nice work and dataset.
I wanna download this dataset for training, but I can't open the OneDrive link, (I have already used vpn), could you give another way to download it like GoogleDrive? Or some solutions to help me get that.

Thanks a lot.

I couldn't find the vanishing points in the dataset. Could you share this information?

It seems that there're a lot of errors relate to path, for example:

1. cannot find .lib
2. joint() argument must be str or bytes, not 'PosixPath'

Could you please tell us exactly the version(3.5, 3.6 or 3.7) of python as well as the system you use?

Thanks

你好！
我想下载数据训练一下模型，但是链接好像并不能下载数据集
不知道方不方便发到百度云上或是其他方式都可以
万分感谢！

作者您好！感谢提供论文和数据集，很有收获。
我们现在想在自己的数据集上进行标注，您方便提比较好的标注工具吗？

I can't access OneDrive, so I can't download data.
Is there any other way to download data?

I wonder if the dataset has camera intrinsic parameter for each image. Thanks :)

Hi,

I would like to ask you what is the exact coordinate system in which the junction points are expressed with their (x,y) coordinates in image space? I see that the data contains both 0 and 500 values, for images with 500xN or Mx500 resolution, practically on one end or the other it is indexing outside of the image boundary, so I'm unsure what would be the exact coordinate system.
If you use OpenCV and indexing from 0, I understand that the top left corner is at [0,0] coordinate, but then where is the point with x=500 located? It's outside of the image space. What would be the correct way to interpret the data? Simply neglect values above 499?

Thank you in advance for the clarification,
Robert

How to annotate new data as wireframe format. Is there any tool to do so? If there is can you please share the name?

Hi Kun Huang and fellow community members,

I am unable to find the mat-files for wireframe evaluation from the OneDrive link. Could someone kindly re-upload the files again or send them to my email [email protected]?

Any help would be greatly appreciated! Thank you!

原先的baidu网盘用不了！

if i change batch size bigger than 2, there is a error like this:
RuntimeError: The size of tensor a (10) must match the size of tensor b (2) at non-singleton dimension 1
how can i fix it?thank you!