Update README.md
Browse files
README.md
CHANGED
|
@@ -59,3 +59,17 @@ inputs = tokenizer(
|
|
| 59 |
scores = model(**inputs).logits.cpu().detach().numpy()
|
| 60 |
print(scores)
|
| 61 |
```
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 59 |
scores = model(**inputs).logits.cpu().detach().numpy()
|
| 60 |
print(scores)
|
| 61 |
```
|
| 62 |
+
|
| 63 |
+
## Training
|
| 64 |
+
|
| 65 |
+
Trained using MSELoss using <query, document> pairs with <grade> as the label.
|
| 66 |
+
|
| 67 |
+
```python
|
| 68 |
+
from sentence_transformers import InputExample
|
| 69 |
+
|
| 70 |
+
train_samples = [
|
| 71 |
+
InputExample(texts=["query 1", "document 1"], label=0.3),
|
| 72 |
+
InputExample(texts=["query 1", "document 2"], label=0.8),
|
| 73 |
+
InputExample(texts=["query 2", "document 2"], label=0.1),
|
| 74 |
+
]
|
| 75 |
+
````
|