From 661f4c473f463aa6cf11165e07e03bbf45a0ff85 Mon Sep 17 00:00:00 2001
From: Julian Risch <julian.risch@deepset.ai>
Date: Wed, 29 Sep 2021 11:28:12 +0200
Subject: [PATCH] Replace FARMRanker with SentenceTransformersRanker

---
 docs/latest/components/ranker.mdx | 16 ++++++++--------
 1 file changed, 8 insertions(+), 8 deletions(-)

diff --git a/docs/latest/components/ranker.mdx b/docs/latest/components/ranker.mdx
index cc72313d7..86c1921ff 100644
--- a/docs/latest/components/ranker.mdx
+++ b/docs/latest/components/ranker.mdx
@@ -17,17 +17,17 @@ Alternatively, [this example](https://github.com/deepset-ai/FARM/blob/master/exa
 
 <div style={{ marginBottom: "3rem" }} />
 
-## FARMRanker
+## SentenceTransformersRanker
 
 ### Description
 
-The FARMRanker consists of a Transformer-based model for document re-ranking using the TextPairClassifier of [FARM](https://github.com/deepset-ai/FARM).
-Given a text pair of query and passage, the TextPairClassifier either predicts label "1" if the pair is similar or label "0" if they are dissimilar (accompanied with a probability).
-While the underlying model can vary (BERT, Roberta, DistilBERT, ...), the interface remains the same.
-With a FARMRanker, you can:
+The SentenceTransformersRanker consists of a Sentence Transformer based pre-trained Cross-Encoder model for Document Re-ranking (https://huggingface.co/cross-encoder).
+Re-Ranking can be used on top of a retriever to boost the performance for document search. This is particularly useful if the retriever has a high recall but is bad in sorting the documents by relevance.
+SentenceTransformerRanker handles Cross-Encoder models that use a single logit as similarity score (https://www.sbert.net/docs/pretrained-models/ce-msmarco.html#usage-with-transformers). This similarity score describes the similarity of the cross-encoded query and document text.
+
+With a SentenceTransformersRanker, you can:
 
 - Directly get predictions (re-ranked version of the supplied list of Document) via predict() if supplying a pre-trained model
-- Take a plain language model (e.g. `bert-base-cased`) and train it for TextPairClassification via train()
 
 <div style={{ marginBottom: "3rem" }} />
 
@@ -36,13 +36,13 @@ With a FARMRanker, you can:
 ```python
 from haystack.document_store import ElasticsearchDocumentStore
 from haystack.retriever import ElasticsearchRetriever
-from haystack.ranker import FARMRanker
+from haystack.ranker import SentenceTransformersRanker
 from haystack import Pipeline
 
 document_store = ElasticsearchDocumentStore()
 ...
 retriever = ElasticsearchRetriever(document_store)
-ranker = FARMRanker(model_name_or_path="saved_models/roberta-base-asnq-binary")
+ranker = SentenceTransformersRanker(model_name_or_path="cross-encoder/ms-marco-MiniLM-L-12-v2")
 ...
 p = Pipeline()
 p.add_node(component=retriever, name="ESRetriever", inputs=["Query"])