Update BM25S model

README.md

@@ -123,9 +123,9 @@ This dataset was created using the following data:
 
 | Statistic | Value |
 | --- | --- |
-| Number of documents | 210801 |
-| Number of tokens | 1247449 |
-| Average tokens per document | 5.92 |
+| Number of documents | 347536 |
+| Number of tokens | 2831191 |
+| Average tokens per document | 8.15 |
 
 ## Parameters
 

corpus.jsonl

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:34a38e542a64bfd8e7e4f75c31ec7f9cae03b26196701e40904bc580cd971150
-size 14491479
+oid sha256:70cbea160b6e2b451f7895ef633be9f65f8f277a4cf719f2d6d1307e07b18944
+size 29870605

data.csc.index.npy

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:99d46b75c6d39fd6b858be4b194d6385f77073896e33955d707086366a09801c
-size 4989924
+oid sha256:10f3d0fd57448dd4dcd62cf39dfc704a7e351a384b3dfce6c7d592c598194e84
+size 11324892

indices.csc.index.npy

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:6d615201000cc91e6961a8b2a102d79cffbfa87610645fbfe414753bcc592d6e
-size 4989924
+oid sha256:2fd3bbbecb5cb9da6f930c947e99d143de87670d8f06f2929fbf217d26703450
+size 11324892

indptr.csc.index.npy

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:ee295a3f6f37a7988013298b018103c98aef8d5861590e4d87391285d40d266c
-size 248616
+oid sha256:52277c0ab954d867d9d502276cc4236c6a4ffa6e81d4af199fd26a4a686589fa
+size 577132

params.index.json

@@ -6,7 +6,7 @@
     "idf_method": "lucene",
     "dtype": "float32",
     "int_dtype": "int32",
-    "num_docs": 210801,
+    "num_docs": 347536,
     "version": "0.2.6",
     "backend": "numpy"
 }

vocab.index.json

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:03d5613c6c043f095d746de405aab145523713820d2756469e919e871e02f5e4
-size 995887
+oid sha256:5262817bfcf340fbc58ca8779dce06fc99093abe768db9ebabb1099bbccb0774
+size 2436979