Datasets:
Update README.md
Browse files
README.md
CHANGED
|
@@ -32,6 +32,27 @@ configs:
|
|
| 32 |
path: data/train-*
|
| 33 |
task_categories:
|
| 34 |
- text-generation
|
|
|
|
| 35 |
size_categories:
|
| 36 |
- 1K<n<10K
|
| 37 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 32 |
path: data/train-*
|
| 33 |
task_categories:
|
| 34 |
- text-generation
|
| 35 |
+
- feature-extraction
|
| 36 |
size_categories:
|
| 37 |
- 1K<n<10K
|
| 38 |
+
license: cc0-1.0
|
| 39 |
+
source_datasets: storytracer/LoC-PD-Books
|
| 40 |
+
language:
|
| 41 |
+
- en
|
| 42 |
+
tags:
|
| 43 |
+
- books
|
| 44 |
+
---
|
| 45 |
+
|
| 46 |
+
|
| 47 |
+
# LoC-PD-Books: preprocessed
|
| 48 |
+
|
| 49 |
+
|
| 50 |
+
This is the `storytracer/LoC-PD-Books` dataset with the following preprocessing steps:
|
| 51 |
+
|
| 52 |
+
- apply [clean-text](https://pypi.org/project/clean-text/) package keeping casing and newlines
|
| 53 |
+
- drop OCR garbled text in first few lines of each example
|
| 54 |
+
- fix (most) 'hard' newlines w/ regex similar to [gutenberg clean](https://huggingface.co/datasets/BEE-spoke-data/gutenberg-en-v1-clean)
|
| 55 |
+
- 'grade' first 512 tokens of each book with [this quantized model](https://huggingface.co/pszemraj/gibberish_detector_onnx-quant-avx512_vnni); keep examples from labels `clean` (all) and `mild gibberish` w/ score 0.9 or higher
|
| 56 |
+
|
| 57 |
+
---
|
| 58 |
+
|