Update README.md

README.md

@@ -89,7 +89,38 @@ On LMEB, reranking models demonstrate a clear advantage, with even the 0.27B Nan
 ![lmeb_emb](./assets/lmeb_emb.jpg)
 
 # Usage
+```python
+import argparse
+
+from kalm_reranker import KaLMReranker
+
+
+def main() -> None:
+    parser = argparse.ArgumentParser()
+    parser.add_argument(
+        "--model",
+        default="KaLM-Embedding/KaLM-Reranker-V1-Large"
+    )
+    parser.add_argument("--device", default=None)
+    args = parser.parse_args()
+
+    reranker = KaLMReranker(args.model, device=args.device)
+    query = "What is the capital of China?"
+    documents = [
+        "The capital of China is Beijing.",
+        "Gravity attracts bodies toward one another.",
+    ]
+    instruction = "Given a query, retrieve documents that answer the query."
 
+    pairs = [(query, document) for document in documents]
+    print("scores:", reranker.predict(pairs, instruction=instruction))
+    print("rankings:", reranker.rank(query, documents, instruction=instruction))
+
+
+if __name__ == "__main__":
+    main()
+
+```
 
 # Citation
 If you find this model useful, please cite our papers.