Dataset Preview
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 7 new columns ({'position', 'bg', 'size', 'bare', 'group', 'hexcode', 'name'}) and 1 missing columns ({'text'}).
This happened while the json dataset builder was generating data using
hf://datasets/ayushmaninbox/artificial-stupidity-emoji/emoji/train/specs.json (at revision 1d8030e8b9fcf3b5a7eab382ef04b385cfe916f0), ['hf://datasets/ayushmaninbox/artificial-stupidity-emoji@1d8030e8b9fcf3b5a7eab382ef04b385cfe916f0/emoji/train/captions.json', 'hf://datasets/ayushmaninbox/artificial-stupidity-emoji@1d8030e8b9fcf3b5a7eab382ef04b385cfe916f0/emoji/train/specs.json', 'hf://datasets/ayushmaninbox/artificial-stupidity-emoji@1d8030e8b9fcf3b5a7eab382ef04b385cfe916f0/emoji300/train/captions.json', 'hf://datasets/ayushmaninbox/artificial-stupidity-emoji@1d8030e8b9fcf3b5a7eab382ef04b385cfe916f0/emoji300/train/specs.json'], ['hf://datasets/ayushmaninbox/artificial-stupidity-emoji@1d8030e8b9fcf3b5a7eab382ef04b385cfe916f0/emoji/train/captions.json', 'hf://datasets/ayushmaninbox/artificial-stupidity-emoji@1d8030e8b9fcf3b5a7eab382ef04b385cfe916f0/emoji/train/specs.json', 'hf://datasets/ayushmaninbox/artificial-stupidity-emoji@1d8030e8b9fcf3b5a7eab382ef04b385cfe916f0/emoji300/train/captions.json', 'hf://datasets/ayushmaninbox/artificial-stupidity-emoji@1d8030e8b9fcf3b5a7eab382ef04b385cfe916f0/emoji300/train/specs.json']
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/python3.14/site-packages/datasets/builder.py", line 1848, in _prepare_split_single
writer.write_table(table)
~~~~~~~~~~~~~~~~~~^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/arrow_writer.py", line 765, in write_table
self._write_table(pa_table, writer_batch_size=writer_batch_size)
~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/arrow_writer.py", line 773, in _write_table
pa_table = table_cast(pa_table, self._schema)
File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2378, in table_cast
return cast_table_to_schema(table, schema)
File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2306, in cast_table_to_schema
raise CastError(
...<3 lines>...
)
datasets.table.CastError: Couldn't cast
name: string
hexcode: string
group: string
position: string
bg: string
size: string
bare: bool
to
{'text': Value('string')}
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 1369, in compute_config_parquet_and_info_response
parquet_operations, partial, estimated_dataset_info = stream_convert_to_parquet(
~~~~~~~~~~~~~~~~~~~~~~~~~^
builder, max_dataset_size_bytes=max_dataset_size_bytes
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
)
^
File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 948, in stream_convert_to_parquet
builder._prepare_split(split_generator=splits_generators[split], file_format="parquet")
~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1694, in _prepare_split
for job_id, done, content in self._prepare_split_single(
~~~~~~~~~~~~~~~~~~~~~~~~~~^
gen_kwargs=gen_kwargs, job_id=job_id, **_prepare_split_args
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
):
^
File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1850, in _prepare_split_single
raise DatasetGenerationCastError.from_cast_error(
...<4 lines>...
)
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 7 new columns ({'position', 'bg', 'size', 'bare', 'group', 'hexcode', 'name'}) and 1 missing columns ({'text'}).
This happened while the json dataset builder was generating data using
hf://datasets/ayushmaninbox/artificial-stupidity-emoji/emoji/train/specs.json (at revision 1d8030e8b9fcf3b5a7eab382ef04b385cfe916f0), ['hf://datasets/ayushmaninbox/artificial-stupidity-emoji@1d8030e8b9fcf3b5a7eab382ef04b385cfe916f0/emoji/train/captions.json', 'hf://datasets/ayushmaninbox/artificial-stupidity-emoji@1d8030e8b9fcf3b5a7eab382ef04b385cfe916f0/emoji/train/specs.json', 'hf://datasets/ayushmaninbox/artificial-stupidity-emoji@1d8030e8b9fcf3b5a7eab382ef04b385cfe916f0/emoji300/train/captions.json', 'hf://datasets/ayushmaninbox/artificial-stupidity-emoji@1d8030e8b9fcf3b5a7eab382ef04b385cfe916f0/emoji300/train/specs.json'], ['hf://datasets/ayushmaninbox/artificial-stupidity-emoji@1d8030e8b9fcf3b5a7eab382ef04b385cfe916f0/emoji/train/captions.json', 'hf://datasets/ayushmaninbox/artificial-stupidity-emoji@1d8030e8b9fcf3b5a7eab382ef04b385cfe916f0/emoji/train/specs.json', 'hf://datasets/ayushmaninbox/artificial-stupidity-emoji@1d8030e8b9fcf3b5a7eab382ef04b385cfe916f0/emoji300/train/captions.json', 'hf://datasets/ayushmaninbox/artificial-stupidity-emoji@1d8030e8b9fcf3b5a7eab382ef04b385cfe916f0/emoji300/train/specs.json']
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.
text string |
|---|
slightly frowning face |
a small left arrow curving right in the left on a teal background |
up left arrow |
full moon face |
a small snow capped mountain in the right on a teal background |
keycap 6 |
a small record button in the bottom on a navy background |
gloves |
clockwise vertical arrows |
last quarter moon face |
a large hourglass not done in the bottom right on a navy background |
a medium sunrise in the top right on a cream background |
eagle |
a large owl in the right on a cream background |
triangular ruler |
a medium red paper lantern in the left on a cream background |
a large red question mark in the bottom on a white background |
ice cream |
a large fire in the right on a white background |
magnifying glass tilted left |
a large bell in the center on a white background |
japanese post office |
a large ok button in the bottom on a grey background |
water buffalo |
female sign |
yellow square |
a large soft ice cream in the center on a black background |
a large ear of corn in the bottom right on a black background |
razor |
chocolate bar |
a medium bacon in the bottom on a navy background |
koala |
a small crocodile in the top right on a white background |
flute |
a small railway track in the top right on a grey background |
a large kissing face with closed eyes in the top right on a navy background |
a small otter in the center on a grey background |
a large christmas tree in the bottom left on a grey background |
saxophone |
trolleybus |
a small railway car in the center on a grey background |
a medium pill in the top right on a white background |
a small bowling in the bottom on a white background |
a small check mark in the top left on a navy background |
a medium ribbon in the right on a grey background |
a medium 3rd place medal in the center on a black background |
a small necktie in the bottom left on a black background |
confused face |
light rail |
hushed face |
a large mending heart in the bottom left on a cream background |
a large cut of meat in the left on a black background |
ewe |
a medium hut in the bottom right on a teal background |
a medium triangular ruler in the top left on a black background |
antenna bars |
a small blue heart in the right on a white background |
a small level slider in the bottom right on a teal background |
ice |
a medium globe with meridians in the bottom on a cream background |
ear of corn |
airplane departure |
a large heart decoration in the bottom left on a navy background |
moai |
a large notebook in the bottom left on a grey background |
rolling on the floor laughing |
a small black small square in the bottom left on a black background |
a small waning gibbous moon in the bottom left on a navy background |
sparkles |
a medium hundred points in the right on a cream background |
a medium curry rice in the bottom on a cream background |
herb |
a small baseball in the top left on a cream background |
a medium package in the left on a white background |
dove |
a large saluting face in the top right on a teal background |
bear |
a medium cooked rice in the top left on a white background |
a small record button in the bottom on a cream background |
red envelope |
a small non potable water in the top right on a white background |
a large cyclone in the bottom on a white background |
a small mans shoe in the bottom right on a navy background |
a large custard in the top on a cream background |
a small yin yang in the bottom left on a white background |
a medium sunrise in the right on a white background |
a medium heavy equals sign in the bottom right on a grey background |
paw prints |
slightly frowning face |
chair |
world map |
a medium convenience store in the top right on a cream background |
a small non potable water in the center on a navy background |
pot of food |
a large e mail in the top left on a black background |
a small nest with eggs in the top left on a navy background |
a large speedboat in the left on a teal background |
old key |
lipstick |
spiral calendar |
End of preview.
Artificial Stupidity — Emoji Corpus
The training corpus for AS-I: 45,000 captioned 64×64 renders built from OpenMoji artwork.
Two splits of the same construction:
| Glyphs | Samples per glyph | |
|---|---|---|
emoji/ |
1,254 | 36 |
emoji300/ |
300 | 150 |
Training identical models on each is what produces the coverage-vs-detail result on the model card.
Caption grammar
Deliberately closed, so that prompt adherence is measurable rather than a matter of opinion — every caption is assembled from known parts, so a benchmark can read the parts back off a generated image and score them exactly:
<name>
a <size> <name> in the <position> on a <bg> background
positions 9 (top left ... bottom right)
backgrounds 6 (white, black, navy, grey, cream, teal)
sizes 3 (small, medium, large)
Example captions:
red heart
a small pizza in the top left on a navy background
a large grinning face in the bottom right on a cream background
Contents
| File | What |
|---|---|
catalog.json |
the glyphs: hexcode, name, group |
train/latents.pt |
VAE-encoded latents, ready for a diffusion prior |
train/captions.json |
one caption per sample |
train/specs.json |
ground truth per sample, for scoring |
Latents are float16, shape (N, 4, 16, 16), already multiplied by the VAE's
scale factor so a prior sees ~unit-variance inputs.
Reproducing the pixels
Images are not redistributed here — OpenMoji artwork is CC BY-SA 4.0 and is better fetched from source. One command rebuilds them byte for byte:
python data/emoji.py --n 45000 --size 64 --seed 1337
Attribution
Artwork: OpenMoji, CC BY-SA 4.0.
- Downloads last month
- 31