Datasets:

bbrothers
/

avspeech-metadata

ArXiv:

Dataset Preview

Split (2)

train

The full dataset viewer is not available (click to read why). Only showing a preview of the rows.

The dataset generation failed because of a cast error

Error code:   DatasetGenerationCastError
Exception:    DatasetGenerationCastError
Message:      An error occurred while generating the dataset

All the data files must have the same columns, but at some point there are 5 new columns ({'0.849219', '0.305556', '111.240000', 'u5MPyrRJPmc', '108.240000'}) and 5 missing columns ({'233.266000', '0.780469', '0.670833', 'CJoOwXcjhds', '239.367000'}).

This happened while the csv dataset builder was generating data using

hf://datasets/bbrothers/avspeech-metadata/avspeech_test.csv (at revision c51a0db620e40bb0552c0d3fc1f13d68e93e5f95)

Please either edit the data files to have matching columns, or separate them into different configurations (see docs at https://hf.co/docs/hub/datasets-manual-configuration#multiple-configurations)
Traceback:    Traceback (most recent call last):
                File "/usr/local/lib/python3.12/site-packages/datasets/builder.py", line 1831, in _prepare_split_single
                  writer.write_table(table)
                File "/usr/local/lib/python3.12/site-packages/datasets/arrow_writer.py", line 714, in write_table
                  pa_table = table_cast(pa_table, self._schema)
                             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/table.py", line 2272, in table_cast
                  return cast_table_to_schema(table, schema)
                         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/table.py", line 2218, in cast_table_to_schema
                  raise CastError(
              datasets.table.CastError: Couldn't cast
              u5MPyrRJPmc: string
              108.240000: double
              111.240000: double
              0.849219: double
              0.305556: double
              -- schema metadata --
              pandas: '{"index_columns": [{"kind": "range", "name": null, "start": 0, "' + 861
              to
              {'CJoOwXcjhds': Value('string'), '233.266000': Value('float64'), '239.367000': Value('float64'), '0.780469': Value('float64'), '0.670833': Value('float64')}
              because column names don't match
              
              During handling of the above exception, another exception occurred:
              
              Traceback (most recent call last):
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 1455, in compute_config_parquet_and_info_response
                  parquet_operations = convert_to_parquet(builder)
                                       ^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 1054, in convert_to_parquet
                  builder.download_and_prepare(
                File "/usr/local/lib/python3.12/site-packages/datasets/builder.py", line 894, in download_and_prepare
                  self._download_and_prepare(
                File "/usr/local/lib/python3.12/site-packages/datasets/builder.py", line 970, in _download_and_prepare
                  self._prepare_split(split_generator, **prepare_split_kwargs)
                File "/usr/local/lib/python3.12/site-packages/datasets/builder.py", line 1702, in _prepare_split
                  for job_id, done, content in self._prepare_split_single(
                                               ^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/builder.py", line 1833, in _prepare_split_single
                  raise DatasetGenerationCastError.from_cast_error(
              datasets.exceptions.DatasetGenerationCastError: An error occurred while generating the dataset
              
              All the data files must have the same columns, but at some point there are 5 new columns ({'0.849219', '0.305556', '111.240000', 'u5MPyrRJPmc', '108.240000'}) and 5 missing columns ({'233.266000', '0.780469', '0.670833', 'CJoOwXcjhds', '239.367000'}).
              
              This happened while the csv dataset builder was generating data using
              
              hf://datasets/bbrothers/avspeech-metadata/avspeech_test.csv (at revision c51a0db620e40bb0552c0d3fc1f13d68e93e5f95)
              
              Please either edit the data files to have matching columns, or separate them into different configurations (see docs at https://hf.co/docs/hub/datasets-manual-configuration#multiple-configurations)

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.

CJoOwXcjhds string	233.266000 float64	239.367000 float64	0.780469 float64	0.670833 float64
AvWWVOgaMlk	90	93.566667	0.586719	0.311111
Y8HMIm8mdns	171.607767	174.607767	0.505729	0.240741
akwvpAiLFk0	144.68	150	0.698438	0.288889
Swss72CHSWg	90.023267	97.2972	0.230729	0.20463
ymD5uLlLc0g	36.033	40.9	0.341667	0.475926
DuWE-CQDlEk	211.266667	221.7	0.788281	0.401389
uCGKDCWxqxo	16.56	30	0.420833	0.482407
-A9gdf3j2xo	295.165	298.165	0.507812	0.233333
labiHToR5nk	266.52	269.52	0.522656	0.243056
xbgfxIc-nbs	116.149367	119.986533	0.504167	0.45
QoQF8N5ZsQA	240.006433	244.961389	0.450781	0.358333
307DK9nGQhw	64.097367	73.006267	0.327604	0.17963
5qy9Ujv9XdM	61.494767	65.331933	0.557813	0.392593
UBL2Vowiulk	30.113422	33.616922	0.725	0.393056
LcyrfLT2tto	95.5985	98.5985	0.471094	0.409722
sujFCXbYkMo	30	34.466667	0.528906	0.477778
R0u9E8GsUXk	114.466667	118.166667	0.879687	0.829167
bNxD_breZy8	198.734	201.734	0.539062	0.240278
AQDWwktBhaw	30.03	40.807433	0.469531	0.293056
Dtn8xZ3BiGY	114.31	119.828	0.489063	0.316667
Cy9SUMj5wGY	16.133333	30	0.789062	0.598611
8nQBG5hvjpk	283.286	286.286	0.792188	0.348611
rCp8Jae81KU	257.6	260.6	0.711979	0.374074
PmD-LzPS2rg	282.5823	285.5853	0.371875	0.425
BrCcDt6GNkk	281.333333	284.466667	0.55	0.244444
IrXrbrZWflA	203.169	209.976	0.416146	0.271296
512K2S3De-A	282.6824	288.855233	0.514844	0.338889
01qWxISqaHg	87	90	0.46875	0.303977
2f32XSMYlDk	73.440033	77.844433	0.150521	0.163889
lcClO5lHEjA	120.086633	124.991533	0.439063	0.348611
q1doqKlHRuY	18.9	23.333333	0.102083	0.834259
tdTXVU5wN8I	188.32	199.28	0.492188	0.421296
h-fSfAFufCo	153.9538	160.226733	0.496354	0.378704
srwckJKdeS0	174.36	177.44	0.51875	0.702778
DIWf1t-HzwI	74.207467	77.210467	0.471875	0.405556
YXcVkIEMGds	295.08	300	0.390104	0.198148
BsUzOhJ9WGU	53.053	59.993	0.497396	0.45
JjduaMIoKvI	266.9697	269.9697	0.476562	0.366667
IlnHVjvBDU0	275.76	288.16	0.475	0.304167
zFH0QbS-l-w	287.153533	291.557933	0.53125	0.418056
h5wT_c4fQ1o	168.201367	179.9798	0.525521	0.27037
3AsPqH3QaQM	273.439833	282.949333	0.524219	0.259722
aaEA__Js2u0	270	273.266667	0.869531	0.7875
7rYeSDHS0U0	120	134.28	0.485156	0.276389
qpEzCs23PWE	134.100633	144.110633	0.501042	0.303704
8E2UlNrLNmk	50.721	53.721	0.4625	0.323148
BjvtZkHWExY	79.370958	89.506083	0.314063	0.197222
qzM4wshoqGs	91.36	94.36	0.555729	0.348148
871zAw-g1ZM	120.566667	124.133333	0.526563	0.386111
JpJoybtabbU	186.866667	190.566667	0.254167	0.465741
lUdymhI3Zl4	203	208.4	0.520833	0.403704
RdUVaYI3bmg	120.522522	127.861611	0.517708	0.09537
Uu1xVo0CF5o	106.5064	119.986533	0.552604	0.285185
WfpZPDqNNg0	260.433	269.999	0.723958	0.248148
G7xm-5aDZyg	114.52	120	0.486458	0.234259
jdshBkVfjrA	293.852	299.975	0.433594	0.476389
3gcWAZSNi2E	30.997633	38.204833	0.598958	0.317593
Wl3HSpsiIb4	229.646089	235.360122	0.627344	0.226389
yeqK6kqoIYk	221.2	227.68	0.306792	0.295833
YWgXhe7JYp4	144.602789	149.899756	0.835417	0.151852
ReXQGb2k3fo	11.166667	22.2	0.417187	0.429167
2RvWHWhyx1w	56.993267	59.993267	0.527344	0.375
AOoqrXx5BNU	164.3	177.267	0.541146	0.544444
_8K1hWkirLo	90	96.64	0.494271	0.300926
cPUBnjqIaXI	189.933333	195.466667	0.797656	0.443056
342Pxxa7n8Q	253.2	256.56	0.308854	0.371296
umcJyBaatBs	105.533333	119.466667	0.607031	0.408333
QYnLgIsR3bc	180.5804	184.150633	0.492969	0.277778
ohd_xOV6zW4	56.993	59.993	0.515104	0.469444
DxpQmBfA6vM	83.086	86.086	0.314583	0.442593
xwxbJkXRJHw	247.64	254.12	0.455469	0.201389
Gqt1A6O6UTk	66.3	69.366	0.759375	0.181944
YOySUCOJUtQ	127.994533	141.107633	0.566406	0.240278
ft3fl0x3gFo	232.966	240	0.520312	0.445833
7DBdAnTuw5c	61.394667	66.524789	0.503125	0.242593
ibPOxQ7XYPk	30.16	37.44	0.478125	0.263889
ausPEm5ZWQE	76.916667	90	0.39604	0.286111
a2iQ7kB5b6s	136.28	140.76	0.517188	0.446296
WddCvVatDlo	240	251.3	0.6375	0.558333
04BgTYq4Ckk	273.25	284.042	0.539844	0.452778
wBDD5wTG7P0	200.866	207.766	0.432031	0.201389
6LapKTptu8w	162.033	172.799	0.517188	0.216667
gX8qtrFaLs4	120.9238	123.9238	0.496354	0.306481
XXtkzCkzA64	106.92	119.92	0.608333	0.313889
TH0xt8XIvVs	120.007	134.352	0.549219	0.295833
H2VCPO0isFQ	31.489789	35.744044	0.817708	0.127778
aVP0dc3dvtA	143.41	149.984	0.485938	0.27963
5fnzANZN-O4	120.125	135.058333	0.558333	0.220833
alHB2f34oRI	115.482033	119.586133	0.568229	0.280556
98iIikUQgWQ	233.666	239.989	0.522656	0.376389
81jh1rIVC0g	257.96	263.08	0.5	0.249074
UVDnhj-jZl0	106.2	110.333333	0.875	0.854167
awqKt7frvJI	166.36	169.36	0.549219	0.25
RfP2dOHPej8	232.498933	239.872967	0.710156	0.347222
GtM2sM6r3So	124.257467	131.9318	0.483594	0.341667
k3NzaNrdALo	74.066667	77.166667	0.699219	0.358333
vK-snIInirc	138.1	149.566667	0.825521	0.4375
r3N3LCHqjdI	164.430933	176.860022	0.457292	0.268519
gOWL5FwU4c4	37.68	42.8	0.496868	0.354167
MZnQ3eZuUAE	241.96	245.24	0.651563	0.441667

End of preview.

YAML Metadata Warning: empty or missing yaml metadata in repo card (https://huggingface.co/docs/hub/datasets-cards)

AVSpeech Metadata Files

This repository contains the metadata CSV files for the AVSpeech dataset by Google Research.

Dataset Description

AVSpeech is a large-scale audio-visual speech dataset containing over 290,000 video segments from YouTube, designed for audio-visual speech recognition and lip reading research.

Files

avspeech_train.csv (128 MB) - Training set with 2,621,845 video segments from 270k videos
avspeech_test.csv (9 MB) - Test set with video segments from a separate set of 22k videos

CSV Format

Each row contains:

YouTube ID, start_time, end_time, x_coordinate, y_coordinate

Where:

YouTube ID: The YouTube video identifier
start_time: Start time of the segment in seconds
end_time: End time of the segment in seconds
x_coordinate: X coordinate of the speaker's face center (normalized 0.0-1.0, 0.0 = left)
y_coordinate: Y coordinate of the speaker's face center (normalized 0.0-1.0, 0.0 = top)

The train and test sets have disjoint speakers.

Usage

With Hugging Face Hub

from huggingface_hub import hf_hub_download

# Download train CSV
train_csv = hf_hub_download(
    repo_id="bbrothers/avspeech-metadata",
    filename="avspeech_train.csv",
    repo_type="dataset"
)

# Download test CSV
test_csv = hf_hub_download(
    repo_id="bbrothers/avspeech-metadata",
    filename="avspeech_test.csv",
    repo_type="dataset"
)

With our dataset loader

from ml.data.av_speech.dataset import AVSpeechDataset

# Initialize dataset (will auto-download CSVs if needed)
dataset = AVSpeechDataset()

# Download videos
dataset.download(
    splits=['train', 'test'],
    max_videos=100,  # Or None for all videos
    num_workers=4
)

Citation

If you use this dataset, please cite the original AVSpeech paper:

@inproceedings{ephrat2018looking,
  title={Looking to listen at the cocktail party: A speaker-independent audio-visual model for speech separation},
  author={Ephrat, Ariel and Mosseri, Inbar and Lang, Oran and Dekel, Tali and Wilson, Kevin and Hassidim, Avinatan and Freeman, William T and Rubinstein, Michael},
  booktitle={ACM SIGGRAPH 2018},
  year={2018}
}

Notes

This repository only contains the metadata CSV files, not the actual video content
Videos must be downloaded from YouTube using the provided YouTube IDs
Some videos may no longer be available (deleted, private, or geo-blocked)
Estimated total dataset size: ~4500 hours of video

Downloads last month: 63

Paper for bbrothers/avspeech-metadata

Looking to Listen at the Cocktail Party: A Speaker-Independent Audio-Visual Model for Speech Separation

Paper • 1804.03619 • Published Apr 10, 2018