The model can be used for Information Retrieval: Given a query, encode the query will all possible passages (e.g. retrieved with ElasticSearch). Then sort the passages in a decreasing order.
Query and Paragraph.
Logits
Set the Query and Paragraph as an argument.
$ python3 cross_encoder_mmarco.py -q "How many people live in Berlin?" -p "Berlin has a population of 3,520,031 registered inhabitants in an area of 891.82 square kilometers."
$ python3 cross_encoder_mmarco.py -q "How many people live in Berlin?" -p "New York City is famous for the Metropolitan Museum of Art."
$ python3 cross_encoder_mmarco.py -q "ベルリンには何人が住んでいますか？" -p "ベルリンの人口は891.82平方キロメートルの地域に登録された住民が3,520,031人います。"
$ python3 cross_encoder_mmarco.py -q "ベルリンには何人が住んでいますか？" -p "ニューヨーク市はメトロポリタン美術館で有名です。"Output : [array([[10.761541]], dtype=float32)]
Output : [array([[-8.127746]], dtype=float32)]
Output : [array([[9.374646]], dtype=float32)]
Output : [array([[-6.408309]], dtype=float32)]jeffwan/mmarco-mMiniLMv2-L12-H384-v
- PyTorch 2.2.1
- Transformers 4.33.3
ONNX opset = 11
XLMRobertaTokenizer (Same with SentenceTransformer and E5)