axotion
/

polish-reranker-base-ranknet

@@ -11,12 +11,15 @@ base_model:
 <h1 align="center">polish-reranker-base-ranknet</h1>
 This is a Polish text ranking model trained with [RankNet loss](https://icml.cc/Conferences/2015/wp-content/uploads/2015/06/icml_ranking.pdf) on a large dataset of text pairs consisting of 1.4 million queries and 10 million documents.
 The training data included the following parts: 1) The Polish MS MARCO training split (800k queries); 2) The ELI5 dataset translated to Polish (over 500k queries); 3) A collection of Polish medical questions and answers (approximately 100k queries).
 As a teacher model, we employed [unicamp-dl/mt5-13b-mmarco-100k](https://huggingface.co/unicamp-dl/mt5-13b-mmarco-100k), a large multilingual reranker based on the MT5-XXL architecture. As a student model, we choose [Polish RoBERTa](https://huggingface.co/sdadas/polish-roberta-base-v2).
 Unlike more commonly used pointwise losses, which regard each query-document pair independently, the RankNet method computes loss based on queries and pairs of documents. More specifically, the loss is computed based on the relative order of documents sorted by their relevance to the query.
 To train the reranker, we used the teacher model to assess the relevance of the documents extracted in the retrieval stage for each query. We then sorted these documents by the relevance score, obtaining a dataset consisting of queries and ordered lists of 20 documents per query.
 ## Usage (Sentence-Transformers)
 You can use the model like this with [sentence-transformers](https://www.SBERT.net):

 <h1 align="center">polish-reranker-base-ranknet</h1>
+### This repository extends the original repository by providing an ONNX version for optimized inference.
 This is a Polish text ranking model trained with [RankNet loss](https://icml.cc/Conferences/2015/wp-content/uploads/2015/06/icml_ranking.pdf) on a large dataset of text pairs consisting of 1.4 million queries and 10 million documents.
 The training data included the following parts: 1) The Polish MS MARCO training split (800k queries); 2) The ELI5 dataset translated to Polish (over 500k queries); 3) A collection of Polish medical questions and answers (approximately 100k queries).
 As a teacher model, we employed [unicamp-dl/mt5-13b-mmarco-100k](https://huggingface.co/unicamp-dl/mt5-13b-mmarco-100k), a large multilingual reranker based on the MT5-XXL architecture. As a student model, we choose [Polish RoBERTa](https://huggingface.co/sdadas/polish-roberta-base-v2).
 Unlike more commonly used pointwise losses, which regard each query-document pair independently, the RankNet method computes loss based on queries and pairs of documents. More specifically, the loss is computed based on the relative order of documents sorted by their relevance to the query.
 To train the reranker, we used the teacher model to assess the relevance of the documents extracted in the retrieval stage for each query. We then sorted these documents by the relevance score, obtaining a dataset consisting of queries and ordered lists of 20 documents per query.
 ## Usage (Sentence-Transformers)
 You can use the model like this with [sentence-transformers](https://www.SBERT.net):