Update BM25S model

README.md

@@ -123,9 +123,9 @@ This dataset was created using the following data:
 
 | Statistic | Value |
 | --- | --- |
-| Number of documents | 750312 |
-| Number of tokens | 7592215 |
-| Average tokens per document | 10.12 |
+| Number of documents | 712943 |
+| Number of tokens | 7391911 |
+| Average tokens per document | 10.37 |
 
 ## Parameters
 

corpus.jsonl

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2b999015937322493801414a7628aea03373b140623efa2d7e247af03e4eb2b2
-size 73690346
+oid sha256:5869a1f3263444bb6292d30075a29405e88740a06255b9e312569b4339001230
+size 69920920

data.csc.index.npy

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e30a1c11c7cc998cca6c57896b569da0e32ca94e6c59b92d14648706c2d670aa
-size 30368988
+oid sha256:70756fa753a7c110e506386956b0f84358d2b123175f8b2b2459b8c255d86f0a
+size 29567772

indices.csc.index.npy

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5b2f427f0d29bab6a4744d4d14cad7f6205617efb8f2381fa0e82664698e1f92
-size 30368988
+oid sha256:a9d24996451232d9b3f3a3d57eeba29f73ccfef68e3116e6b74c8626bb140d92
+size 29567772

indptr.csc.index.npy

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e01faf6b39cfa30b9e13fb00c062117e7487e6a6e8b4539522a37f6b30591a52
-size 559348
+oid sha256:fcb12500c5b017e78b2f608ddc910f48d876bdeeee3fabad7ebe47262e5e89c8
+size 553400

params.index.json

@@ -6,7 +6,7 @@
     "idf_method": "lucene",
     "dtype": "float32",
     "int_dtype": "int32",
-    "num_docs": 750312,
+    "num_docs": 712943,
     "version": "0.2.7post1",
     "backend": "numpy"
 }

vocab.index.json

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d224cd8071210250b768abb4e03969c446bef1a2d382ad711325ad63c54af4c9
-size 2283889
+oid sha256:45500e26ddca7d47ccec5ef4dcc61036ffaecc5d2ddc99e1a1bab55d7cea10b5
+size 2252514