Update README.md
Browse files
README.md
CHANGED
|
@@ -42,4 +42,26 @@ configs:
|
|
| 42 |
path: data/val-*
|
| 43 |
- split: test
|
| 44 |
path: data/test-*
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 45 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 42 |
path: data/val-*
|
| 43 |
- split: test
|
| 44 |
path: data/test-*
|
| 45 |
+
task_categories:
|
| 46 |
+
- text-generation
|
| 47 |
+
language:
|
| 48 |
+
- ru
|
| 49 |
+
tags:
|
| 50 |
+
- paraphrase-generation
|
| 51 |
+
- paraphrase
|
| 52 |
+
size_categories:
|
| 53 |
+
- 100K<n<1M
|
| 54 |
---
|
| 55 |
+
## Dataset Description
|
| 56 |
+
|
| 57 |
+
The dataset is obtained by filtering [dataset of russian paraphrases by David Dale](https://huggingface.co/datasets/cointegrated/ru-paraphrase-NMT-Leipzig) with automatic metrics.
|
| 58 |
+
The data structure is saved.
|
| 59 |
+
Have been deleted:
|
| 60 |
+
- Paraphrases that have cosine labse similarity with source sentences < 0.75.
|
| 61 |
+
- Paraphrases that are 2.5 times longer than source sentences. (Most of them are looped errors of back translation)
|
| 62 |
+
- Paraphrases that are similar in spelling to the original texts. Paraphrases that have ChrF++ similarity > 0.6 and consist of 15 or more tokens have been deleted.
|
| 63 |
+
|
| 64 |
+
Source dataset contained samples with English paraphrases (errors of back translation). This paraphrases have been translated into Russian.
|
| 65 |
+
|
| 66 |
+
### Supported Tasks and Leaderboards
|
| 67 |
+
The dataset can be used to train and evaluate models for paraphrase generation or (if negative sampling is used) for paraphrase detection.
|