Issue with jalFaizy/detect_chess_pieces when running datasets-cli test

[faizankshaikh]

Link

https://huggingface.co/datasets/jalFaizy/detect_chess_pieces

Description

I am trying to write a appropriate data loader for a custom dataset using this script

When I run the command

$ datasets-cli test "D:\workspace\HF\detect_chess_pieces" --save_infos --all_configs

It gives the following error

Using custom data configuration default
Traceback (most recent call last):
  File "c:\users\faiza\anaconda3\lib\runpy.py", line 194, in _run_module_as_main
    return _run_code(code, main_globals, None,
  File "c:\users\faiza\anaconda3\lib\runpy.py", line 87, in _run_code
    exec(code, run_globals)
  File "C:\Users\faiza\anaconda3\Scripts\datasets-cli.exe\__main__.py", line 7, in <module>
  File "c:\users\faiza\anaconda3\lib\site-packages\datasets\commands\datasets_cli.py", line 39, in main
    service.run()
  File "c:\users\faiza\anaconda3\lib\site-packages\datasets\commands\test.py", line 132, in run
    for j, builder in enumerate(get_builders()):
  File "c:\users\faiza\anaconda3\lib\site-packages\datasets\commands\test.py", line 125, in get_builders
    yield builder_cls(
  File "c:\users\faiza\anaconda3\lib\site-packages\datasets\builder.py", line 1148, in __init__
    super().__init__(*args, **kwargs)
  File "c:\users\faiza\anaconda3\lib\site-packages\datasets\builder.py", line 306, in __init__
    info = self.get_exported_dataset_info()
  File "c:\users\faiza\anaconda3\lib\site-packages\datasets\builder.py", line 405, in get_exported_dataset_info
    return self.get_all_exported_dataset_infos().get(self.config.name, DatasetInfo())
  File "c:\users\faiza\anaconda3\lib\site-packages\datasets\builder.py", line 390, in get_all_exported_dataset_infos
    return DatasetInfosDict.from_directory(cls.get_imported_module_dir())
  File "c:\users\faiza\anaconda3\lib\site-packages\datasets\info.py", line 309, in from_directory
    dataset_infos_dict = {
  File "c:\users\faiza\anaconda3\lib\site-packages\datasets\info.py", line 310, in <dictcomp>
    config_name: DatasetInfo.from_dict(dataset_info_dict)
  File "c:\users\faiza\anaconda3\lib\site-packages\datasets\info.py", line 272, in from_dict
    return cls(**{k: v for k, v in dataset_info_dict.items() if k in field_names})
  File "<string>", line 20, in __init__
  File "c:\users\faiza\anaconda3\lib\site-packages\datasets\info.py", line 160, in __post_init__
    templates = [
  File "c:\users\faiza\anaconda3\lib\site-packages\datasets\info.py", line 161, in <listcomp>
    template if isinstance(template, TaskTemplate) else task_template_from_dict(template)
  File "c:\users\faiza\anaconda3\lib\site-packages\datasets\tasks\__init__.py", line 43, in task_template_from_dict
    return template.from_dict(task_template_dict)
AttributeError: 'NoneType' object has no attribute 'from_dict'

My assumption is that there is some kind of issue in how the "task_templates" are read, because even if I keep them as None, or not include the argument at all, the same error occurs

Owner

Yes

[albertvillanova]

Hi @faizankshaikh

Please note that we have recently launched the Community feature, specifically targeted to create Discussions (about issues/questions/asking-for-help) on each Dataset on the Hub:

• Blog post: https://huggingface.co/blog/community-update
• Docs: https://huggingface.co/docs/hub/repositories-pull-requests-discussions

The Discussion tab for your "jalFaizy/detect_chess_pieces" dataset is here: https://huggingface.co/datasets/jalFaizy/detect_chess_pieces/discussions
You can use it to ask for help by pinging the Datasets maintainers: see our docs here: https://huggingface.co/docs/datasets/master/en/share#ask-for-a-help-and-reviews

I'm transferring this discussion to your Discussion tab and trying to address it: https://huggingface.co/datasets/jalFaizy/detect_chess_pieces/discussions/1

[faizankshaikh]

Thank you @albertvillanova , I will keep that in mind.

Just a quick note - I posted the issue on Github because the dataset viewer suggested me to "open an issue for direct support". Maybe it can be updated with your suggestion

[albertvillanova]

Thank you pointing this out: yes, definitely, we should fix the error message. We are working on this.