prithivida
/

miniReranker_arabic_v1

Text Classification

Transformers.js

passage-retrieval

knowledge-distillation

Inference Endpoints

Model card Files Files and versions

prithivida commited on Aug 15

Commit

c0e99b1

•

1 Parent(s): 8136180

Update README.md

Files changed (1) hide show

README.md +20 -17

README.md CHANGED Viewed

@@ -8,7 +8,7 @@ tags:
   - reranker
   - cross-encoder
   - transformers.js
-  - miniDense
   - passage-retrieval
   - knowledge-distillation
 library_name: transformers
@@ -22,6 +22,9 @@ library_name: transformers
 # Usage and Performance
 ### With Cross Encoders (for quick testing, for production use see next section)
 ```python
@@ -66,22 +69,22 @@ readable_scores
 ```
 ```
-['0.00058645',
- '0.00683665',
- '0.00250791',
- '0.00061849',
- '0.00162472',
- '0.97978389',
- '0.03804804',
- '0.90570509',
- '0.82151824',
- '0.32221058',
- '0.10520715',
- '0.12972574',
- '0.06127196',
- '0.00659027',
- '0.38976011',
- '0.91317886']
 ```
 ### With Flashrank (Save $ and increase speed, No heavy Torch or Transformers dependencies)

   - reranker
   - cross-encoder
   - transformers.js
+  - miniReranker
   - passage-retrieval
   - knowledge-distillation
 library_name: transformers
 # Usage and Performance
+## This model is a Arabic equivalent of [ms-marco-MiniLM-L-12-v2 reranker](https://www.sbert.net/docs/pretrained-models/ce-msmarco.html).
+  - MSMARCO Dev score TBA.
 ### With Cross Encoders (for quick testing, for production use see next section)
 ```python
 ```
 ```
+['0.00058640',
+ '0.00684038',
+ '0.00250885',
+ '0.00061850',
+ '0.00162680',
+ '0.97980452',
+ '0.03813050',
+ '0.90573835',
+ '0.82196653',
+ '0.32288009',
+ '0.10534699',
+ '0.12966448',
+ '0.06130386',
+ '0.00660836',
+ '0.39035776',
+ '0.91324186']
 ```
 ### With Flashrank (Save $ and increase speed, No heavy Torch or Transformers dependencies)