book_id stringlengths 7 22 | split stringclasses 1
value | source stringclasses 4
values | text stringlengths 188 1.75M | spans listlengths 0 1.88k |
|---|---|---|---|---|
I001D528B | train | reviewed_opf | "༄༅། །རང་བྱུང་པདྨའི་སྙིང་ཐིག་སྔོན་(...TRUNCATED) | [
{
"id": "a1d8b208135a463fb3d380b74ee734fc",
"start": 206248,
"end": 207747
}
] |
I0288940D | train | reviewed_opf | "༄༅། །མཉམ་མེད་རྒོད་ཚང་པའི་ཞལ་གདམས་(...TRUNCATED) | [
{
"id": "8f6d413e7e534ec69894405b60c1fcb9",
"start": 29981,
"end": 30072
}
] |
I07ECF0AE | train | reviewed_opf | "༄༅། །བོད་ཀྱི་བརྡ་དག་བྱིས་པའི་ལམ་མ(...TRUNCATED) | [
{
"id": "0cf10473675f4aa087d5bd88bc35d42a",
"start": 4849,
"end": 4985
}
] |
I08405360 | train | reviewed_opf | "༄༅། །སྡེ་བདུན་མདོ་དང་བཅས་པའི་དགོང(...TRUNCATED) | [
{
"id": "25cf49c0dd714862854a171fdf67ccb4",
"start": 346793,
"end": 346880
}
] |
I0FCFA88F | train | reviewed_opf | "དྭགས་བརྒྱུད་སྒྲུབ་པའི་ཐེམ་སྐས་གདམ(...TRUNCATED) | [{"id":"c3fb7d2dc07b4ac59c69f05d836a19e9","start":92142,"end":92156},{"id":"b089221dac9f41af9c3dbb02(...TRUNCATED) |
I100E7DAD | train | reviewed_opf | "དྭགས་བརྒྱུད་སྒྲུབ་པའི་ཐེམ་སྐས་གདམ(...TRUNCATED) | [{"id":"3107d562a46a42899f89af6ed54b60f6","start":188854,"end":189120},{"id":"5c7024c98d774c38b05789(...TRUNCATED) |
I163178A0 | train | reviewed_opf | "༄༅། །སངས་རྒྱས་ཀྱི་སྐུའི་རབ་ཏུ་དབྱ(...TRUNCATED) | [{"id":"1b201eef6028405cad6f67c21ee88dcb","start":6908,"end":6986},{"id":"5c9429a1b790402986c1a2b556(...TRUNCATED) |
I19A08A51 | train | reviewed_opf | "༄༅། །བདེ་གཤེགས་སྙིང་པོའི་བསྟོད་པ་(...TRUNCATED) | [{"id":"075c29fa6b9f453687f938a74dd4bfdb","start":14266,"end":14466},{"id":"fc8109fb97e742969fb3a5e6(...TRUNCATED) |
I1CC06C39 | train | reviewed_opf | "༄༅། །དྲི་ལན་སྐོར་རྨོངས་པའི་མུན་སེ(...TRUNCATED) | [{"id":"bc6029f4757e4963a4de2eb5e1c0a9de","start":248,"end":281},{"id":"7fabc23ff36f4a1c84eb6aaae568(...TRUNCATED) |
I1DCBD5CC | train | reviewed_opf | "༄༅། །ཕྱག་རྒྱ་ཆེན་པོའི་སྨོན་ལམ་གྱི(...TRUNCATED) | [{"id":"dd5649552e10484fb5d04e0267728580","start":3858,"end":3868},{"id":"b3a49ac8590647589822cff924(...TRUNCATED) |
Tibetan Yigchung Combined Annotation Detection (v2.0)
This is the combined dataset (v2.0) assembling reviewed OPF, Lotsawa House, Nalanda Vinaya, and 44 BDRC volumes. Dedup-removed books/passages are left out of Lotsawa, BDRC, and OPF train only.
762 Classical Tibetan books annotated for Yigchung (small-print interlinear annotation), stored as character-offset spans over a flat base text. 29,084 spans across 76,904,236 characters.
Offsets are inclusive [start, end] — text[start:end+1].
Sources
reviewed_opf: 228 bookslotsawa: 469 booksnalanda: 21 booksbdrc: 44 books
Splits
Book-level: a book never appears in more than one split.
The Hub configs use validation instead of val.
| Split | Books | Spans | Text chars | Chars in spans | Spans/100k | Mean span len |
|---|---|---|---|---|---|---|
train |
601 | 22,388 | 60,977,652 | 3.734% | 36.72 | 101.7 |
validation |
84 | 2,637 | 7,456,195 | 3.547% | 35.37 | 100.3 |
test |
77 | 4,059 | 8,470,389 | 4.169% | 47.92 | 87.0 |
Config windowed_w8192_s4914
Derived data. Regenerate it if the tokenizer, window length or overlap changes. Never mix windows from different tokenizers or strides in one run.
Windows are cut from a full untruncated tokenization (content tokens per
window = 8192 minus 2 specials = 8190). The last window of a book is
end-aligned. Hugging Face return_overflowing_tokens is not used.
from datasets import load_dataset
ds = load_dataset("karma689/yigchung_detection_v2", "windowed_w8192_s4914", revision="v2.0")
| Field | Value |
|---|---|
| Tokenizer | jhu-clsp/mmBERT-base |
Window length (max_length) |
8192 |
| Content tokens per window | 8190 |
| Overlap (tokens) | 3276 |
| Measured token step | 4914 |
| Label scheme | 3-class BIO (O / B-YIGCHUNG / I-YIGCHUNG) |
| Gold spans split across windows | 0 |
| Source commit | 3a5ca09d484f6edb2bed6bbdc8ba63e634e06dc8 |
tokenizer.json sha256 |
197d4cc5406ee12cc50c8b5511f2393cc32d9db321545979ce041c1199178356 |
tokenizer_config.json sha256 |
1d2f82c1341a79748e00efe82e67690f99d00b3c2a894f2b23128fd9d3519da3 |
Labels are already {0: O, 1: B-YIGCHUNG, 2: I-YIGCHUNG}.
Window counts
| Split | Books | Windows |
|---|---|---|
| train | 601 | 8,574 |
| validation | 84 | 1,041 |
| test | 77 | 1,189 |
Overlapped tokens are labelled -100 in every window except the first that
covers them, so each content token contributes to the loss exactly once.
Labels
| id | label |
|---|---:|---|
| 0 | O |
| 1 | B-YIGCHUNG |
| 2 | I-YIGCHUNG |
-100 marks special tokens and tokens already labelled by an earlier window.
Columns
| Column | Description |
|---|---|
book_id |
OpenPecha or BDRC/Lotsawa/Nalanda id |
source |
Source corpus (reviewed_opf, lotsawa, nalanda, bdrc) |
window_index |
0-based window within the book |
input_ids / attention_mask |
tokenizer output, no padding |
labels |
3-class BIO, -100 on specials and re-covered tokens |
char_start / char_end |
window coverage in book characters (char_end exclusive) |
own_start / own_end |
owned token range for stitching |
token_start / token_end |
window coverage in book tokens |
spans |
gold spans fully inside this window |
License
Source texts were digitized and made available by the Buddhist Digital Resource Center (BDRC) and Lotsawa House. We gratefully acknowledge BDRC, Tsadra Foundation, and OpenPecha.
- Downloads last month
- 74