Rerank#

Learn how to use rerank models in Xinference.

Introduction#

Given a query and a list of documents, Rerank indexes the documents from most to least semantically relevant to the query. Rerank models in Xinference can be invoked through the Rerank endpoint to rank a list of documents.

Quickstart#

We can try Rerank API out either via cURL, OpenAI Client, or Xinference’s python client:

curl -X 'POST' \
  'http://<XINFERENCE_HOST>:<XINFERENCE_PORT>/v1/rerank' \
  -H 'accept: application/json' \
  -H 'Content-Type: application/json' \
  -d '{
    "model": "<MODEL_UID>",
    "query": "A man is eating pasta.",
    "documents": [
        "A man is eating food.",
        "A man is eating a piece of bread.",
        "The girl is carrying a baby.",
        "A man is riding a horse.",
        "A woman is playing violin."
    ]
  }'

from xinference.client import Client

client = Client("http://<XINFERENCE_HOST>:<XINFERENCE_PORT>")
model = client.get_model(<MODEL_UID>)

query = "A man is eating pasta."
corpus = [
    "A man is eating food.",
    "A man is eating a piece of bread.",
    "The girl is carrying a baby.",
    "A man is riding a horse.",
    "A woman is playing violin."
]
print(model.rerank(corpus, query))

{
    "id": "480dca92-8910-11ee-b76a-c2c8e4cad3f5",
    "results": [{
        "index": 0,
        "relevance_score": 0.9999247789382935,
        "document": "A man is eating food."
    }, {
        "index": 1,
        "relevance_score": 0.2564932405948639,
        "document": "A man is eating a piece of bread."
    }, {
        "index": 3,
        "relevance_score": 0.00003955026841140352,
        "document": "A man is riding a horse."
    }, {
        "index": 2,
        "relevance_score": 0.00003742107219295576,
        "document": "The girl is carrying a baby."
    }, {
        "index": 4,
        "relevance_score": 0.00003739788007806055,
        "document": "A woman is playing violin."
    }]
}

Multimodal reranking with jina-reranker-m0#

jina-reranker-m0 supports text and image reranking with a maximum context length of 10,240 tokens. Text documents use the standard rerank API shown above. To rerank image URLs or local image paths, pass doc_type="image":

model.rerank(
    [
        "https://example.com/document-1.png",
        "https://example.com/document-2.png",
    ],
    "Which document describes a small language model?",
    doc_type="image",
)

For an image query, also pass query_type="image".