Update BM25S model

README.md

@@ -123,9 +123,9 @@ This dataset was created using the following data:
 
 | Statistic | Value |
 | --- | --- |
-| Number of documents | 362841 |
-| Number of tokens | 2975453 |
-| Average tokens per document | 8.2 |
+| Number of documents | 831507 |
+| Number of tokens | 8338070 |
+| Average tokens per document | 10.03 |
 
 ## Parameters
 

corpus.jsonl

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:9f53137b4941dda5acc77bc9aaf71e87674bfdd8c0e210f7da39835929823189
-size 31663181
+oid sha256:16164bf429cfc58ca04d9782582ab594274462760bc2eb0cc776e07d095cc5d5
+size 85381127

data.csc.index.npy

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:b2f9838d0102b28876970161fa72ef842c60e8745dd855a336d95f4c5cab5694
-size 11901940
+oid sha256:9c698db1df7ff3cb81bc33c0e162561368ebbc3dd40cf3bc4e1ea747d1234407
+size 33352408

indices.csc.index.npy

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f8f216e8ad4aa8d8a5842155212809db8f6271b634722a72b23493087b2d8f12
-size 11901940
+oid sha256:f1c9a4c9f1c7423f165ff28fbff534b9cb486281f7aa29246899c9eda84521dc
+size 33352408

indptr.csc.index.npy

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:241b5484c5bb54f67b0b0d04b3e6d3d09a669c44c57f0ca6ed03ea630abcb43d
-size 577168
+oid sha256:03d05dfbd9c6cf91f4b4d3f7047f44c573bd55fadee26ee6a4fbefc25b920fe8
+size 1408676

params.index.json

@@ -6,7 +6,7 @@
     "idf_method": "lucene",
     "dtype": "float32",
     "int_dtype": "int32",
-    "num_docs": 362841,
+    "num_docs": 831507,
     "version": "0.2.6",
     "backend": "numpy"
 }

vocab.index.json

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:b1d934afdf72d46521d9543421eda21214c8dcbf975c20f20841c0e63e13ce8e
-size 2437143
+oid sha256:736514f073d877676ddb3bb40b61362d0d9974b4e0826c84ed13ccd47e59b414
+size 6301373