Can't download vots2023

[ngoctuan1]

Hi u, I have installed vote-toolkit 0.6.1. But I can't load the VOT 2023 dataset. It fails when it has reached 16%. Traceback (most recent call last): File "/home/server1-ailab/miniconda3/envs/tuan_toolkit/lib/python3.8/site-packages/vot/utilities/cli.py", line 409, in main do_workspace(args) File "/home/server1-ailab/miniconda3/envs/tuan_toolkit/lib/python3.8/site-packages/vot/utilities/cli.py", line 168, in do_workspace Workspace.initialize(config.workspace, default_config, download=not config.nodownload) File "/home/server1-ailab/miniconda3/envs/tuan_toolkit/lib/python3.8/site-packages/vot/workspace/__init__.py", line 103, in initialize Workspace.download_dataset(dataset, dataset_directory) File "/home/server1-ailab/miniconda3/envs/tuan_toolkit/lib/python3.8/site-packages/vot/workspace/__init__.py", line 118, in download_dataset download_dataset(dataset, directory) File "/home/server1-ailab/miniconda3/envs/tuan_toolkit/lib/python3.8/site-packages/vot/dataset/__init__.py", line 584, in download_dataset download_dataset_meta(url, path) File "/home/server1-ailab/miniconda3/envs/tuan_toolkit/lib/python3.8/site-packages/vot/dataset/vot.py", line 200, in download_dataset_meta raise DatasetException("Unable do download channel bundle") vot.dataset.DatasetException: Unable do download channel bundle

[lukacu]

Hi, we have checked again the downloading process and it completes fine. Does it fail at the same sequence every time?

[ngoctuan1]

I uninstalled vot-toolkit and reinstalled it. It's version 0.6.1. And I re-downloaded 'vot initialize vots2023 --workspace data_2023'. When the process reaches 16%, there will be an error[image: image.png] [image: image.png] [image: image.png]

Vào CN, 14 thg 5, 2023 vào lúc 19:36 Luka Čehovin Zajc < @.***> đã viết:

Hi, we have checked again the downloading process and it completes fine. Does it fail at the same sequence every time?

— Reply to this email directly, view it on GitHub https://github.com/votchallenge/toolkit/issues/75#issuecomment-1546890242, or unsubscribe https://github.com/notifications/unsubscribe-auth/ARHGYF3HRETVP4SRGPNU34LXGDGTHANCNFSM6AAAAAAX7LMJGQ . You are receiving this because you authored the thread.Message ID: @.***>

[franktpmvu]

same stop in 16%, it says

Downloading sequence dataset "VOTS2023 Challenge" with 144 sequences.
 Downloading          |███████████████▍                                                                                 |  16% [11:59<1:03:05]
File is not a zip file
Traceback (most recent call last):
  File "/opt/conda/lib/python3.7/site-packages/vot/utilities/cli.py", line 409, in main
    do_workspace(args)
  File "/opt/conda/lib/python3.7/site-packages/vot/utilities/cli.py", line 168, in do_workspace
    Workspace.initialize(config.workspace, default_config, download=not config.nodownload)
  File "/opt/conda/lib/python3.7/site-packages/vot/workspace/__init__.py", line 103, in initialize
    Workspace.download_dataset(dataset, dataset_directory)
  File "/opt/conda/lib/python3.7/site-packages/vot/workspace/__init__.py", line 118, in download_dataset
    download_dataset(dataset, directory)
  File "/opt/conda/lib/python3.7/site-packages/vot/dataset/__init__.py", line 584, in download_dataset
    download_dataset_meta(url, path)
  File "/opt/conda/lib/python3.7/site-packages/vot/dataset/vot.py", line 198, in download_dataset_meta
    download_uncompress(channel_url, channel_directory)
  File "/opt/conda/lib/python3.7/site-packages/vot/utilities/net.py", line 129, in download_uncompress
    extract_files(tmp_file, path)
  File "/opt/conda/lib/python3.7/site-packages/vot/utilities/__init__.py", line 169, in extract_files
    with ZipFile(file=archive) as zip_file:
  File "/opt/conda/lib/python3.7/zipfile.py", line 1258, in __init__
    self._RealGetContents()
  File "/opt/conda/lib/python3.7/zipfile.py", line 1325, in _RealGetContents
    raise BadZipFile("File is not a zip file")
zipfile.BadZipFile: File is not a zip file

i think it stopped on file 'carchase'

[lukacu]

Ok, so the carchase sequence is apparently quite large, the largest in the dataset. It did not cause problems for us when testing, but it may be that the chance of getting an error is much higher. We are investigating how to improve the download, expect a solution or at least a bypass soon (today).

[JornyWan]

Ok, so the carchase sequence is apparently quite large, the largest in the dataset. It did not cause problems for us when testing, but it may be that the chance of getting an error is much higher. We are investigating how to improve the download, expect a solution or at least a bypass soon (today).

hi lukacu, i think you can improve the download process by fixing the sequence named "carchase" which seems cause the error : File is not a zip file.

[lukacu]

We have released version 0.6.2 of the toolkit that should improve the download process, try it out and report back.

[JornyWan]

We have released version 0.6.2 of the toolkit that should improve the download process, try it out and report back.

thank you for your fast update, i tried to vot initialize but when it download at 13%, there is an error: Downloading sequence dataset "VOTS2023 Challenge" with 144 sequences (total 26.0GiB). Downloading |████████▊ | 13% [22:16<1:30:55] Error when downloading file, trying to resume download Downloading |█████████▎ | 14% [31:54<7:01:48]

it seems that the download could continue but i'm worried about that there might be some file lost after finish the download

--------update below------- Even after reporting several errors the download finally finished, showing that squirrel was not downloaded successfully, but I added the downloaded squirrel before the update, this time I got the data of carchase, after I evaluate and analysis I found the same as my previous answer below, rusults=[], please tell me why, using is the python_ncc.py you provided. Thanks again!

[JornyWan]

And I would like to know that the results I get when I evaluate the tracker with python_ncc in your integretion after using my vots2023 dataset earlier are empty: results = [] "results": { "baseline": { "parameters": { "realtime": null, "noise": null, "inject": null, "transformers": [], "analyses": [], "repetitions": 1, "early_stop": true, "multiobject": true }, "type": "vot.experiment.multirun.UnsupervisedExperiment", "results": [] } }

and only the json file can output the data contents and rusults above, or if the output is an html file, it can't output result: VOT report Toolkit version: 0.6.1 Created: 2023-05-16T22:08:17.203861

[lukacu]

We are mixing two things now and it is not a good thing in case of issue discussion. So the dataset download is now trying to resume downloads, files are large and connections break. If the download fails, it will still try to download the dataset again before running the tracker and it will skip sequences that are already downloaded.

The other issue is regarding results: please read participation instructions, this year we are running the challenge without public groundtruth, you have to submit results to our server (URL will be available) and you get back the scores. This means that you cannot really run analysis locally.