Dataset Viewer
Auto-converted to Parquet Duplicate
text_id
int64
chunk_id
int64
pixel_values
list
grapheme_token_ids
list
llama_token_ids
list
text
string
0
0
[[255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,25(...TRUNCATED)
[1,6,13,21,4,6,58,111,4,91,55,17,111,89,62,22,4,49,28,8,4,6,13,21,4,6,58,111,4,6,27,4,14,70,5,51,82,(...TRUNCATED)
[1,1094,314,626,1789,410,6772,262,6352,638,302,3761,28899,408,314,626,1789,385,29884,4516,4159,29891(...TRUNCATED)
"Asam amino protéinogenik nyaéta asam amino anu diruntuykeun sacara biosintétik jadi protéin nal(...TRUNCATED)
1
0
[[255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,25(...TRUNCATED)
[ 1, 102, 44, 53, 4, 39, 91, 55, 17, 111, 89, 62, 22, 39, 4, 153, 26, 11, 4, 39, 209, 71, 5, 4, 91, 55, 17, 5, 39, 2 ]
[ 1, 476, 687, 481, 376, 771, 6772, 262, 6352, 638, 29908, 298, 442, 1099, 376, 1460, 347, 348, 410, 6772, 262, 29908, 2 ]
Kecap "protéinogenik" hartina "nyieun protéin"
2
0
[[255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,25(...TRUNCATED)
[1,14,4,13,297,4,16,606,22,4,47,70,53,4,6,27,4,144,12,4,10,49,142,6,5,9,4,6,61,4,7,45,45,7,4,6,13,21(...TRUNCATED)
[1,4671,269,557,332,611,4415,2679,298,381,786,385,29884,1737,375,413,1384,29745,273,29892,263,3761,2(...TRUNCATED)
"Di sakur mahluk hirup anu geus kanyahoan, aya 22 asam amino anu dikode sacara genetik (protéinogen(...TRUNCATED)
3
0
[[255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,25(...TRUNCATED)
[1,13,23,52,22,11,4,26,4,28,8,9,4,6,13,21,4,6,58,111,4,111,5,59,91,55,17,111,89,62,22,4,126,4,6,13,2(...TRUNCATED)
[1,11775,284,638,1056,19538,904,941,29892,408,314,626,1789,1661,29899,771,6772,262,6352,638,611,2988(...TRUNCATED)
"Sabalikna ti éta, asam amino non-protéinogenik mah asam amino anu teu ngaruntuy kana protéin ata(...TRUNCATED)
4
0
[[255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,25(...TRUNCATED)
[1,6,27,4,19,5,184,41,4,65,6,13,11,4,23,73,136,4,26,11,4,122,14,40,10,18,4,107,12,59,118,5,265,18,4,(...TRUNCATED)
[1,530,29884,282,4182,5338,24003,1648,6411,2679,279,260,1099,878,666,638,6840,926,29899,3286,3333,29(...TRUNCATED)
Anu pandeuri biasana balukar tina modipikasi pos-translasi kana protéin
5
0
[[255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,25(...TRUNCATED)
[1,13,23,23,20,36,4,6,13,21,4,6,58,111,4,111,5,59,91,55,17,111,89,62,22,4,14,70,5,51,82,35,5,4,10,11(...TRUNCATED)
[1,11775,370,279,25613,408,314,626,1789,1661,29899,771,6772,262,6352,638,4516,4159,29891,446,348,413(...TRUNCATED)
"Sababaraha asam amino non-protéinogenik diruntuykeun kana péptida non-ribosomal anu disintésis k(...TRUNCATED)
6
0
[[255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,25(...TRUNCATED)
[1,366,4,71,10,41,32,15,4,67,4,91,10,41,32,15,4,65,13,4,231,25,53,35,5,4,57,114,111,18,12,55,17,5,4,(...TRUNCATED)
[1,17966,321,2679,1306,327,1444,686,410,29895,1306,327,2652,29874,14265,481,446,348,269,4498,359,391(...TRUNCATED)
"Boh eukariot jeung prokariot bisa nyelapkeun selenosistéin kana protéinna ngaliwatan runtuyan nuk(...TRUNCATED)
7
0
[[255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,25(...TRUNCATED)
[1,14,11,4,13,23,23,20,36,4,91,10,41,32,15,4,127,8,111,89,62,22,9,4,66,60,5,4,31,6,97,4,32,123,4,65,(...TRUNCATED)
[1,360,1099,15296,370,279,25613,410,29895,1306,327,11510,273,6352,638,29892,413,28480,501,10051,3671(...TRUNCATED)
Dina sababaraha prokariot métanogenik, kodon UAG ogé bisa ditarjamahkeun jadi pirolisin
8
0
[[255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,25(...TRUNCATED)
[1,14,11,4,71,10,41,32,15,9,4,6,61,4,31,297,4,7,45,34,7,4,6,13,21,4,6,58,111,4,91,55,17,111,89,62,22(...TRUNCATED)
[1,360,1099,321,2679,1306,327,29892,263,3761,18293,332,29871,29906,29896,408,314,626,1789,410,6772,2(...TRUNCATED)
"Dina eukariot, aya ukur 21 asam amino protéinogenik, 20 kode genetik baku, ditambah selenosistéin(...TRUNCATED)
9
0
[[255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,255,25(...TRUNCATED)
[1,6,29,22,4,16,27,13,4,65,13,4,209,5,55,18,12,4,7,34,45,7,11,9,4,6,11,107,5,4,27,4,7,69,7,4,6,27,4,(...TRUNCATED)
[1,22886,557,767,11326,2652,29874,7098,524,743,275,29871,29896,29906,1056,29892,385,481,265,4948,298(...TRUNCATED)
"Awak manusa bisa nyintésis 12na, anapon nu 9 anu katelah asam amino ésénsial kudu ngala tina dah(...TRUNCATED)
End of preview. Expand in Data Studio

Sundanese PixelGPT Dataset

This dataset contains preprocessed Sundanese text data for training PixelGPT models.

Dataset Statistics

  • Language: Sundanese (sunda)
  • Total samples: 294,756
  • Train samples: 293,933
  • Test samples: 823

Tokenizers

  • Grapheme tokenizer: izzako/sunda-llama-tokenizer
  • LLaMA tokenizer: ernie-research/DualGPT

Features

  • text_id: Document identifier
  • chunk_id: Chunk identifier within document
  • pixel_values: Rendered pixel representation of aksara text
  • grapheme_token_ids: Token IDs from grapheme-based tokenizer (with EOS token)
  • llama_token_ids: Token IDs from LLaMA tokenizer (with EOS token)
  • text: Original text chunk

Renderer Configuration

  • Renderer path: ../../renderers/m4_renderer

Usage

from datasets import load_dataset

# Load the dataset
dataset = load_dataset("izzako/sundanese-pixelgpt")

# Access train and test splits
train_data = dataset['train']
test_data = dataset['test']

# Example: Get first sample
sample = train_data[0]
print(sample['text'])

Citation

If you use this dataset, please cite:

@dataset{sundanese_pixelgpt,
  title = {Sundanese PixelGPT Dataset},
  author = {Musa Izzanardi Wijanarko},
  year = {2025},
  publisher = {Hugging Face},
  url = {https://huggingface.co/datasets/izzako/sundanese-pixelgpt}
}
Downloads last month
116