Datasets:
Card in the standard format: BDRC names, data source and preparation
Browse files
README.md
CHANGED
|
@@ -1,5 +1,13 @@
|
|
| 1 |
---
|
| 2 |
license: other
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 3 |
language:
|
| 4 |
- bo
|
| 5 |
task_categories:
|
|
@@ -41,17 +49,21 @@ configs:
|
|
| 41 |
path: windowed_w8192_s4579_bioe_reviewed/test/*.parquet
|
| 42 |
---
|
| 43 |
|
| 44 |
-
#
|
|
|
|
|
|
|
| 45 |
|
| 46 |
**254 classical Tibetan books** with **23,095 human-reviewed quotation spans**, split by book into train / validation / test.
|
| 47 |
|
| 48 |
Character offsets are **inclusive on both ends**: the span is `text[start:end+1]`.
|
| 49 |
|
| 50 |
-
##
|
|
|
|
|
|
|
| 51 |
|
| 52 |
Classical Buddhist commentaries digitized with support from the Tsadra Foundation and OpenPecha, annotated in two batches: an old batch (125 books, `Quotation` layer) and a new batch (195 books, `Citation` layer). Both layers mark the same thing — one text quoting a scripture, a named teacher, or a stock phrase — so they are merged into one label.
|
| 53 |
|
| 54 |
-
|
| 55 |
|
| 56 |
1. **Human review.** 398 books reviewed; 78 rejected whole for bad boundaries; 254 boundaries repaired in accepted books.
|
| 57 |
2. **Span merging.** The new batch often splits one quotation across lines. Pieces separated only by 0–2 spaces, newlines, or a shad (and not already closed) were joined: 28,892 → 23,476 spans.
|
|
@@ -100,7 +112,7 @@ Books are **stratified by quote type** (`TEXT` / `PERSON` / `MISC`), so each spl
|
|
| 100 |
|
| 101 |
```python
|
| 102 |
from datasets import load_dataset
|
| 103 |
-
ds = load_dataset("
|
| 104 |
```
|
| 105 |
|
| 106 |
## Windowed configs (for mmBERT training)
|
|
@@ -117,7 +129,7 @@ All quote types collapse to one `QUOTE` label. Load the parquet files directly (
|
|
| 117 |
```python
|
| 118 |
from datasets import load_dataset
|
| 119 |
cfg = "windowed_w8192_s4579" # or windowed_w8192_s4579_bioe
|
| 120 |
-
base = f"hf://datasets/
|
| 121 |
ds = load_dataset("parquet", data_files={
|
| 122 |
s: f"{base}/{s}/*.parquet" for s in ["train", "validation", "test"]
|
| 123 |
})
|
|
@@ -148,7 +160,7 @@ The BIO config is also pinned at tag `w1.0`.
|
|
| 148 |
title = {Tibetan Quotation Detection Benchmark Dataset},
|
| 149 |
author = {karma689},
|
| 150 |
year = {2026},
|
| 151 |
-
url = {https://huggingface.co/datasets/
|
| 152 |
}
|
| 153 |
```
|
| 154 |
|
|
@@ -189,7 +201,7 @@ Two new configs, both test-only:
|
|
| 189 |
|
| 190 |
```python
|
| 191 |
from datasets import load_dataset
|
| 192 |
-
test = load_dataset("
|
| 193 |
data_files={"test": "windowed_w8192_s4579_bioe_reviewed/test/*.parquet"},
|
| 194 |
revision="w2.1")["test"]
|
| 195 |
```
|
|
|
|
| 1 |
---
|
| 2 |
license: other
|
| 3 |
+
pretty_name: Bo-Quotation-Dataset
|
| 4 |
+
tags:
|
| 5 |
+
- tibetan
|
| 6 |
+
- quotation
|
| 7 |
+
- bio
|
| 8 |
+
- bioe
|
| 9 |
+
- mmbert
|
| 10 |
+
- openpecha
|
| 11 |
language:
|
| 12 |
- bo
|
| 13 |
task_categories:
|
|
|
|
| 49 |
path: windowed_w8192_s4579_bioe_reviewed/test/*.parquet
|
| 50 |
---
|
| 51 |
|
| 52 |
+
# Bo-Quotation-Dataset
|
| 53 |
+
|
| 54 |
+
Benchmark and training data for finding **quotation spans** in classical Tibetan commentaries. It is the training data of [`BDRC/Bo-Quotation-Detection`](https://huggingface.co/BDRC/Bo-Quotation-Detection).
|
| 55 |
|
| 56 |
**254 classical Tibetan books** with **23,095 human-reviewed quotation spans**, split by book into train / validation / test.
|
| 57 |
|
| 58 |
Character offsets are **inclusive on both ends**: the span is `text[start:end+1]`.
|
| 59 |
|
| 60 |
+
## Data source and preparation
|
| 61 |
+
|
| 62 |
+
**Source.**
|
| 63 |
|
| 64 |
Classical Buddhist commentaries digitized with support from the Tsadra Foundation and OpenPecha, annotated in two batches: an old batch (125 books, `Quotation` layer) and a new batch (195 books, `Citation` layer). Both layers mark the same thing — one text quoting a scripture, a named teacher, or a stock phrase — so they are merged into one label.
|
| 65 |
|
| 66 |
+
**Cleaning**, in order:
|
| 67 |
|
| 68 |
1. **Human review.** 398 books reviewed; 78 rejected whole for bad boundaries; 254 boundaries repaired in accepted books.
|
| 69 |
2. **Span merging.** The new batch often splits one quotation across lines. Pieces separated only by 0–2 spaces, newlines, or a shad (and not already closed) were joined: 28,892 → 23,476 spans.
|
|
|
|
| 112 |
|
| 113 |
```python
|
| 114 |
from datasets import load_dataset
|
| 115 |
+
ds = load_dataset("BDRC/Bo-Quotation-Dataset")
|
| 116 |
```
|
| 117 |
|
| 118 |
## Windowed configs (for mmBERT training)
|
|
|
|
| 129 |
```python
|
| 130 |
from datasets import load_dataset
|
| 131 |
cfg = "windowed_w8192_s4579" # or windowed_w8192_s4579_bioe
|
| 132 |
+
base = f"hf://datasets/BDRC/Bo-Quotation-Dataset/{cfg}"
|
| 133 |
ds = load_dataset("parquet", data_files={
|
| 134 |
s: f"{base}/{s}/*.parquet" for s in ["train", "validation", "test"]
|
| 135 |
})
|
|
|
|
| 160 |
title = {Tibetan Quotation Detection Benchmark Dataset},
|
| 161 |
author = {karma689},
|
| 162 |
year = {2026},
|
| 163 |
+
url = {https://huggingface.co/datasets/BDRC/Bo-Quotation-Dataset}
|
| 164 |
}
|
| 165 |
```
|
| 166 |
|
|
|
|
| 201 |
|
| 202 |
```python
|
| 203 |
from datasets import load_dataset
|
| 204 |
+
test = load_dataset("BDRC/Bo-Quotation-Dataset",
|
| 205 |
data_files={"test": "windowed_w8192_s4579_bioe_reviewed/test/*.parquet"},
|
| 206 |
revision="w2.1")["test"]
|
| 207 |
```
|