update model

kermitt2 · Jul 18, 2024 · 3780482 · 3780482
1 parent 7bcecd2
commit 3780482
Show file tree

Hide file tree

Showing 5 changed files with 1,201 additions and 28 deletions.
diff --git a/grobid-home/models/name-address-BidLSTM_CRF_FEATURES/config.json b/grobid-home/models/name-address-BidLSTM_CRF_FEATURES/config.json
@@ -2,7 +2,7 @@
     "model_name": "name-address-BidLSTM_CRF_FEATURES",
     "architecture": "BidLSTM_CRF_FEATURES",
     "embeddings_name": "glove-840B",
-    "char_vocab_size": 175,
+    "char_vocab_size": 177,
     "case_vocab_size": 8,
     "char_embedding_size": 25,
     "num_char_lstm_units": 25,

diff --git a/grobid-home/models/name-address-BidLSTM_CRF_FEATURES/model_weights.hdf5 b/grobid-home/models/name-address-BidLSTM_CRF_FEATURES/model_weights.hdf5
diff --git a/grobid-home/models/name-address-BidLSTM_CRF_FEATURES/preprocessor.json b/grobid-home/models/name-address-BidLSTM_CRF_FEATURES/preprocessor.json
@@ -156,32 +156,34 @@
         "\u0107": 146,
         "\u010c": 147,
         "\u010d": 148,
-        "\u0119": 149,
-        "\u0131": 150,
-        "\u0141": 151,
-        "\u0142": 152,
-        "\u0144": 153,
-        "\u0151": 154,
-        "\u0158": 155,
-        "\u015b": 156,
-        "\u015e": 157,
-        "\u0160": 158,
-        "\u0161": 159,
-        "\u017a": 160,
-        "\u017b": 161,
-        "\u017c": 162,
-        "\u017e": 163,
-        "\u02c7": 164,
-        "\u02d8": 165,
-        "\u02dd": 166,
-        "\u03f0": 167,
-        "\u2019": 168,
-        "\u2020": 169,
-        "\u2021": 170,
-        "\u2022": 171,
-        "\u20ac": 172,
-        "\u22a5": 173,
-        "\u2424": 174
+        "\u0111": 149,
+        "\u0119": 150,
+        "\u011f": 151,
+        "\u0131": 152,
+        "\u0141": 153,
+        "\u0142": 154,
+        "\u0144": 155,
+        "\u0151": 156,
+        "\u0158": 157,
+        "\u015b": 158,
+        "\u015e": 159,
+        "\u0160": 160,
+        "\u0161": 161,
+        "\u017a": 162,
+        "\u017b": 163,
+        "\u017c": 164,
+        "\u017e": 165,
+        "\u02c7": 166,
+        "\u02d8": 167,
+        "\u02dd": 168,
+        "\u03f0": 169,
+        "\u2019": 170,
+        "\u2020": 171,
+        "\u2021": 172,
+        "\u2022": 173,
+        "\u20ac": 174,
+        "\u22a5": 175,
+        "\u2424": 176
     },
     "vocab_tag": {
         "<PAD>": 0,