michaelfeil / infinity

Infinity is a high-throughput, low-latency REST API for serving text-embeddings, reranking models and clip
https://michaelfeil.github.io/infinity/
MIT License
1.32k stars 97 forks source link

cannot use rerank (BAAI/bge-base-en-v1.5) #106

Closed BlazJurisic closed 7 months ago

BlazJurisic commented 7 months ago

"message": "InternalServerError: the loaded moded cannot fullyfill rerank.options are {'embed'} inherited from model_class=<class 'infinity_emb.transformer.embedder.sentence_transformer.SentenceTransformerPatched'>",

michaelfeil commented 7 months ago

bge-base-en is not an reranker model. It does not have pytorch weights for classification / a head.

use this model for that: https://huggingface.co/BAAI/bge-reranker-base