Update README.md

README.md

@@ -59,3 +59,17 @@ inputs = tokenizer(
 scores = model(**inputs).logits.cpu().detach().numpy()
 print(scores)
 ```
+
+## Training
+
+Trained using MSELoss using <query, document> pairs with <grade> as the label.
+
+```python
+from sentence_transformers import InputExample
+
+train_samples = [
+    InputExample(texts=["query 1", "document 1"], label=0.3),
+    InputExample(texts=["query 1", "document 2"], label=0.8),
+    InputExample(texts=["query 2", "document 2"], label=0.1),
+]
+````