The dataset viewer is not available for this subset.
Exception: SplitsNotFoundError
Message: The split names could not be parsed from the dataset config.
Traceback: Traceback (most recent call last):
File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 290, in _generate_tables
pa_table = paj.read_json(
io.BytesIO(batch), read_options=paj.ReadOptions(block_size=block_size)
)
File "pyarrow/_json.pyx", line 342, in pyarrow._json.read_json
File "pyarrow/error.pxi", line 155, in pyarrow.lib.pyarrow_internal_check_status
return check_status(status)
File "pyarrow/error.pxi", line 92, in pyarrow.lib.check_status
raise convert_status(status)
pyarrow.lib.ArrowInvalid: JSON parse error: Column() changed from object to string in row 0
During handling of the above exception, another exception occurred:
Traceback (most recent call last):
File "/usr/local/lib/python3.14/site-packages/datasets/inspect.py", line 286, in get_dataset_config_info
for split_generator in builder._split_generators(
~~~~~~~~~~~~~~~~~~~~~~~~~^
StreamingDownloadManager(base_path=builder.base_path, download_config=download_config)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
)
^
File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 101, in _split_generators
pa_table = next(iter(self._generate_tables(**splits[0].gen_kwargs, allow_full_read=False)))[1]
~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 304, in _generate_tables
batch = json_encode_fields_in_json_lines(original_batch, json_field_paths)
File "/usr/local/lib/python3.14/site-packages/datasets/utils/json.py", line 111, in json_encode_fields_in_json_lines
examples = [ujson_loads(line) for line in original_batch.splitlines()]
~~~~~~~~~~~^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/utils/json.py", line 20, in ujson_loads
return pd.io.json.ujson_loads(*args, **kwargs)
~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^
ValueError: Expected object or value
The above exception was the direct cause of the following exception:
Traceback (most recent call last):
File "/src/services/worker/src/worker/job_runners/config/split_names.py", line 66, in compute_split_names_from_streaming_response
for split in get_dataset_split_names(
~~~~~~~~~~~~~~~~~~~~~~~^
path=dataset,
^^^^^^^^^^^^^
config_name=config,
^^^^^^^^^^^^^^^^^^^
token=hf_token,
^^^^^^^^^^^^^^^
)
^
File "/usr/local/lib/python3.14/site-packages/datasets/inspect.py", line 340, in get_dataset_split_names
info = get_dataset_config_info(
path,
...<6 lines>...
**config_kwargs,
)
File "/usr/local/lib/python3.14/site-packages/datasets/inspect.py", line 291, in get_dataset_config_info
raise SplitsNotFoundError("The split names could not be parsed from the dataset config.") from err
datasets.inspect.SplitsNotFoundError: The split names could not be parsed from the dataset config.Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.
UniSign: Israeli Sign Language (ISL) Video & Skeleton Dataset
Dataset Summary
The UniSign dataset is a collection of Israeli Sign Language (ISL) gesture videos and their corresponding pre-extracted skeletal keypoints. It is designed to train models for automated, two-way ISL translation. The dual structure of this dataset supports end-to-end computer vision pipelines as well as direct training of spatial-temporal graph convolutional networks (ST-GCN) using the pre-calculated MediaPipe coordinates.
The video data was sourced from the Institute for the Advancement of Deaf Persons in Israel (Sela) and has been segmented into individual words and phrases.
Languages
- Hebrew (he): The primary spoken/written language mapping for the ISL signs.
- English (en): Provided as an English translation for each sign.
- Israeli Sign Language (ISL): The visual language captured in the videos.
Dataset Structure
Data Instances
Each instance in the dataset consists of a paired video clip (.mp4 or .mov) and a corresponding .json file containing the extracted skeletal landmarks (via MediaPipe).
Data Fields (from words.xlsx)
id: The numerical index of the word.Original Hebrew Folder: The original Hebrew word representing the sign.Suggested ASCII Base: A filesystem-safe ASCII transliteration (e.g., aba, ach). This serves as the class label and folder name.English Translation: The English meaning of the sign.
Directory Structure
The dataset is organized into parallel directories for videos and skeletons to ensure easy pairing via file names.
dataset/
├── words.xlsx
├── videos/
│ ├── aba/
│ │ ├── aba_var01_p01_t01.mp4
│ │ └── ...
│ └── ach/
│ ├── ach_var01_p01_t01.mp4
│ └── ...
└── skeletons/
├── aba/
│ ├── aba_var01_p01_t01.json
│ └── ...
└── ach/
├── ach_var01_p01_t01.json
└── ...
Skeletal Data Format (.json)
The JSON files contain the keypoint coordinates extracted from the videos. (Note: You can add a brief description here of exactly what your JSON contains, e.g., "Contains 33 pose landmarks, 21 left-hand landmarks, and 21 right-hand landmarks extracted using MediaPipe Holistic.")
Dataset Creation
Source Data and Augmentation
The foundational video for each sign was originally sourced and segmented from the official YouTube channel of the Institute for the Advancement of Deaf Persons in Israel (Sela). To ensure robustness, capture natural variations in signing, and increase the volume of training data, the dataset was significantly expanded through custom recordings.
For every original sign, two additional participants recorded 7 distinct video variations. This augmentation provides multiple angles, execution speeds, and slight personal stylistic differences.
- Primary Source Channel: dpiisela YouTube Playlists
- Custom Recordings: 14 additional video variations per sign (7 unique takes per participant across 2 distinct participants).
Annotations
Annotations, indexing, and translations between Hebrew, English, and the ASCII filesystem mappings were structured into the words.xlsx index.
Acknowledgements
Special thanks to the creators of the original video content at the Institute for the Advancement of Deaf Persons in Israel. This dataset is compiled as part of the UniSign academic project.
- Downloads last month
- 14