Update BM25S model

README.md

@@ -123,9 +123,9 @@ This dataset was created using the following data:
 
 | Statistic | Value |
 | --- | --- |
-| Number of documents | 347536 |
-| Number of tokens | 2831191 |
-| Average tokens per document | 8.15 |
+| Number of documents | 362841 |
+| Number of tokens | 2975453 |
+| Average tokens per document | 8.2 |
 
 ## Parameters
 

corpus.jsonl

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:70cbea160b6e2b451f7895ef633be9f65f8f277a4cf719f2d6d1307e07b18944
-size 29870605
+oid sha256:9f53137b4941dda5acc77bc9aaf71e87674bfdd8c0e210f7da39835929823189
+size 31663181

data.csc.index.npy

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:10f3d0fd57448dd4dcd62cf39dfc704a7e351a384b3dfce6c7d592c598194e84
-size 11324892
+oid sha256:b2f9838d0102b28876970161fa72ef842c60e8745dd855a336d95f4c5cab5694
+size 11901940

indices.csc.index.npy

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2fd3bbbecb5cb9da6f930c947e99d143de87670d8f06f2929fbf217d26703450
-size 11324892
+oid sha256:f8f216e8ad4aa8d8a5842155212809db8f6271b634722a72b23493087b2d8f12
+size 11901940

indptr.csc.index.npy

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:52277c0ab954d867d9d502276cc4236c6a4ffa6e81d4af199fd26a4a686589fa
-size 577132
+oid sha256:241b5484c5bb54f67b0b0d04b3e6d3d09a669c44c57f0ca6ed03ea630abcb43d
+size 577168

params.index.json

@@ -6,7 +6,7 @@
     "idf_method": "lucene",
     "dtype": "float32",
     "int_dtype": "int32",
-    "num_docs": 347536,
+    "num_docs": 362841,
     "version": "0.2.6",
     "backend": "numpy"
 }

vocab.index.json

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5262817bfcf340fbc58ca8779dce06fc99093abe768db9ebabb1099bbccb0774
-size 2436979
+oid sha256:b1d934afdf72d46521d9543421eda21214c8dcbf975c20f20841c0e63e13ce8e
+size 2437143