Update README.md
Browse files
README.md
CHANGED
|
@@ -154,14 +154,14 @@ Despite originally being intended for Natural Language Inference (NLI), this dat
|
|
| 154 |
|
| 155 |
### `pair-score` subset
|
| 156 |
|
| 157 |
-
* Columns: "sentence1", "sentence2", "
|
| 158 |
* Column types: `str`, `str`, `float`
|
| 159 |
* Examples:
|
| 160 |
```python
|
| 161 |
{
|
| 162 |
'sentence1': 'A person on a horse jumps over a broken down airplane.',
|
| 163 |
'sentence2': 'A person is training his horse for a competition.',
|
| 164 |
-
'
|
| 165 |
}
|
| 166 |
```
|
| 167 |
* Collection strategy: Taking the `pair-class` subset and remapping "entailment", "neutral" and "contradiction" to 1.0, 0.5 and 0.0, respectively.
|
|
|
|
| 154 |
|
| 155 |
### `pair-score` subset
|
| 156 |
|
| 157 |
+
* Columns: "sentence1", "sentence2", "score"
|
| 158 |
* Column types: `str`, `str`, `float`
|
| 159 |
* Examples:
|
| 160 |
```python
|
| 161 |
{
|
| 162 |
'sentence1': 'A person on a horse jumps over a broken down airplane.',
|
| 163 |
'sentence2': 'A person is training his horse for a competition.',
|
| 164 |
+
'score': 0.5,
|
| 165 |
}
|
| 166 |
```
|
| 167 |
* Collection strategy: Taking the `pair-class` subset and remapping "entailment", "neutral" and "contradiction" to 1.0, 0.5 and 0.0, respectively.
|