Update README.md
Browse files
README.md
CHANGED
@@ -81,7 +81,7 @@ It can give false negatives and false positives on occasion (see Training Result
|
|
81 |
|
82 |
## Training and evaluation data
|
83 |
|
84 |
-
|
85 |
|
86 |
## Usage
|
87 |
```python
|
|
|
81 |
|
82 |
## Training and evaluation data
|
83 |
|
84 |
+
The model is trained using the Natural Questions dataset, with answers that have been refined using GPT-3.5 Turbo. It is evaluated using a number of metrics including BLEU, ROUGE, METEOR, and cosine similarity.
|
85 |
|
86 |
## Usage
|
87 |
```python
|