model improved

README.md

@@ -16,7 +16,7 @@ pipeline_tag: "token-classification"
 
 ## Model Description
 
-This is a BERT model pre-trained with [UD_German-HDT](https://github.com/UniversalDependencies/UD_German-HDT) for POS-tagging and dependency-parsing, derived from [bert-base-german-cased](https://huggingface.co/bert-base-german-cased). Every word is tagged by [UPOS](https://universaldependencies.org/u/pos/) (Universal Part-Of-Speech).
+This is a BERT model pre-trained with [UD_German-HDT](https://github.com/UniversalDependencies/UD_German-HDT) for POS-tagging and dependency-parsing, derived from [gbert-base](https://huggingface.co/deepset/gbert-base). Every word is tagged by [UPOS](https://universaldependencies.org/u/pos/) (Universal Part-Of-Speech).
 
 ## How to Use
 

config.json

@@ -21,56 +21,44 @@
     "10": "B-AUX",
     "11": "B-CCONJ",
     "12": "B-DET",
-    "13": "B-DET+NOUN",
-    "14": "B-INTJ",
-    "15": "B-NOUN",
-    "16": "B-NOUN+PROPN",
-    "17": "B-NUM",
-    "18": "B-PRON",
-    "19": "B-PROPN",
-    "20": "B-PROPN+PROPN",
-    "21": "B-PUNCT",
-    "22": "B-SCONJ",
-    "23": "B-VERB",
-    "24": "B-X",
-    "25": "CCONJ",
-    "26": "CCONJ+NOUN",
-    "27": "DET",
-    "28": "DET+NOUN",
-    "29": "I-ADJ",
-    "30": "I-ADP",
-    "31": "I-ADP+DET",
-    "32": "I-ADV",
-    "33": "I-AUX",
-    "34": "I-CCONJ",
-    "35": "I-DET",
-    "36": "I-DET+NOUN",
-    "37": "I-INTJ",
-    "38": "I-NOUN",
-    "39": "I-NOUN+PROPN",
-    "40": "I-NUM",
-    "41": "I-PRON",
-    "42": "I-PROPN",
-    "43": "I-PROPN+PROPN",
-    "44": "I-PUNCT",
-    "45": "I-SCONJ",
-    "46": "I-VERB",
-    "47": "I-X",
-    "48": "INTJ",
-    "49": "NOUN",
-    "50": "NOUN+PROPN",
-    "51": "NOUN+X",
-    "52": "NUM",
-    "53": "PART",
-    "54": "PRON",
-    "55": "PROPN",
-    "56": "PROPN+PROPN",
-    "57": "PROPN+X",
-    "58": "PUNCT",
-    "59": "SCONJ",
-    "60": "SYM",
-    "61": "VERB",
-    "62": "X"
+    "13": "B-INTJ",
+    "14": "B-NOUN",
+    "15": "B-NUM",
+    "16": "B-PRON",
+    "17": "B-PROPN",
+    "18": "B-PUNCT",
+    "19": "B-SCONJ",
+    "20": "B-VERB",
+    "21": "B-X",
+    "22": "CCONJ",
+    "23": "DET",
+    "24": "I-ADJ",
+    "25": "I-ADP",
+    "26": "I-ADP+DET",
+    "27": "I-ADV",
+    "28": "I-AUX",
+    "29": "I-CCONJ",
+    "30": "I-DET",
+    "31": "I-INTJ",
+    "32": "I-NOUN",
+    "33": "I-NUM",
+    "34": "I-PRON",
+    "35": "I-PROPN",
+    "36": "I-PUNCT",
+    "37": "I-SCONJ",
+    "38": "I-VERB",
+    "39": "I-X",
+    "40": "INTJ",
+    "41": "NOUN",
+    "42": "NUM",
+    "43": "PART",
+    "44": "PRON",
+    "45": "PROPN",
+    "46": "PUNCT",
+    "47": "SCONJ",
+    "48": "SYM",
+    "49": "VERB",
+    "50": "X"
   },
   "initializer_range": 0.02,
   "intermediate_size": 3072,
@@ -88,56 +76,44 @@
     "B-AUX": 10,
     "B-CCONJ": 11,
     "B-DET": 12,
-    "B-DET+NOUN": 13,
-    "B-INTJ": 14,
-    "B-NOUN": 15,
-    "B-NOUN+PROPN": 16,
-    "B-NUM": 17,
-    "B-PRON": 18,
-    "B-PROPN": 19,
-    "B-PROPN+PROPN": 20,
-    "B-PUNCT": 21,
-    "B-SCONJ": 22,
-    "B-VERB": 23,
-    "B-X": 24,
-    "CCONJ": 25,
-    "CCONJ+NOUN": 26,
-    "DET": 27,
-    "DET+NOUN": 28,
-    "I-ADJ": 29,
-    "I-ADP": 30,
-    "I-ADP+DET": 31,
-    "I-ADV": 32,
-    "I-AUX": 33,
-    "I-CCONJ": 34,
-    "I-DET": 35,
-    "I-DET+NOUN": 36,
-    "I-INTJ": 37,
-    "I-NOUN": 38,
-    "I-NOUN+PROPN": 39,
-    "I-NUM": 40,
-    "I-PRON": 41,
-    "I-PROPN": 42,
-    "I-PROPN+PROPN": 43,
-    "I-PUNCT": 44,
-    "I-SCONJ": 45,
-    "I-VERB": 46,
-    "I-X": 47,
-    "INTJ": 48,
-    "NOUN": 49,
-    "NOUN+PROPN": 50,
-    "NOUN+X": 51,
-    "NUM": 52,
-    "PART": 53,
-    "PRON": 54,
-    "PROPN": 55,
-    "PROPN+PROPN": 56,
-    "PROPN+X": 57,
-    "PUNCT": 58,
-    "SCONJ": 59,
-    "SYM": 60,
-    "VERB": 61,
-    "X": 62
+    "B-INTJ": 13,
+    "B-NOUN": 14,
+    "B-NUM": 15,
+    "B-PRON": 16,
+    "B-PROPN": 17,
+    "B-PUNCT": 18,
+    "B-SCONJ": 19,
+    "B-VERB": 20,
+    "B-X": 21,
+    "CCONJ": 22,
+    "DET": 23,
+    "I-ADJ": 24,
+    "I-ADP": 25,
+    "I-ADP+DET": 26,
+    "I-ADV": 27,
+    "I-AUX": 28,
+    "I-CCONJ": 29,
+    "I-DET": 30,
+    "I-INTJ": 31,
+    "I-NOUN": 32,
+    "I-NUM": 33,
+    "I-PRON": 34,
+    "I-PROPN": 35,
+    "I-PUNCT": 36,
+    "I-SCONJ": 37,
+    "I-VERB": 38,
+    "I-X": 39,
+    "INTJ": 40,
+    "NOUN": 41,
+    "NUM": 42,
+    "PART": 43,
+    "PRON": 44,
+    "PROPN": 45,
+    "PUNCT": 46,
+    "SCONJ": 47,
+    "SYM": 48,
+    "VERB": 49,
+    "X": 50
   },
   "layer_norm_eps": 1e-12,
   "max_position_embeddings": 512,
@@ -152,10 +128,6 @@
         "selbst\u00e4ndigeDenken\u00b8": [
           "selbst\u00e4ndige",
           "Denken\u00b8"
-        ],
-        "\u00f6ffentlichenInternetcaf\u00e9s": [
-          "\u00f6ffentlichen",
-          "Internetcaf\u00e9s"
         ]
       },
       "ADP+DET": {
@@ -291,58 +263,6 @@
           "\u00fcber",
           "das"
         ]
-      },
-      "CCONJ+NOUN": {
-        "sowieInternetcaf\u00e9s": [
-          "sowie",
-          "Internetcaf\u00e9s"
-        ]
-      },
-      "DET+NOUN": {
-        "dieCr\u00e8me": [
-          "die",
-          "Cr\u00e8me"
-        ],
-        "dieFort\u00e9-Familie": [
-          "die",
-          "Fort\u00e9-Familie"
-        ]
-      },
-      "NOUN+PROPN": {
-        "Highend-ModellCli\u00e9": [
-          "Highend-Modell",
-          "Cli\u00e9"
-        ],
-        "McKinsey-BeraterRen\u00e9": [
-          "McKinsey-Berater",
-          "Ren\u00e9"
-        ],
-        "NachfolgemodellCli\u00e9": [
-          "Nachfolgemodell",
-          "Cli\u00e9"
-        ]
-      },
-      "PROPN+PROPN": {
-        "AlexanderArtop\u00e9": [
-          "Alexander",
-          "Artop\u00e9"
-        ],
-        "ClausS\u00f8rensen": [
-          "Claus",
-          "S\u00f8rensen"
-        ],
-        "Jean-LouisGass\u00e9e": [
-          "Jean-Louis",
-          "Gass\u00e9e"
-        ],
-        "JoelleRichardi\u00e9re": [
-          "Joelle",
-          "Richardi\u00e9re"
-        ],
-        "LeD\u00e9aut": [
-          "Le",
-          "D\u00e9aut"
-        ]
       }
     }
   },
@@ -351,5 +271,5 @@
   "transformers_version": "4.17.0",
   "type_vocab_size": 2,
   "use_cache": true,
-  "vocab_size": 30000
+  "vocab_size": 31102
 }

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:22c187760763df4a0ba59d50560fd2464f373b3e771ffd63b4bf2a81430eae04
-size 434244299
+oid sha256:4b142e370aff8d696daa9032fe62347bc15e9a5536b7486ff390b7e949806b8e
+size 437592708

supar.model

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a6767137a559c0cda3f40dacaa05190e965713588b2e3d5aea8a355f830e9cad
-size 513838054
+oid sha256:2209e84685d556258b7ce1e2a62e03fcb4cdb0f651fb32cb37836c74feea9cc9
+size 517221736

tokenizer_config.json

@@ -1 +1 @@
-{"do_lower_case": false, "unk_token": "[UNK]", "sep_token": "[SEP]", "pad_token": "[PAD]", "cls_token": "[CLS]", "mask_token": "[MASK]", "tokenize_chinese_chars": true, "strip_accents": null, "model_max_length": 512, "tokenizer_class": "BertTokenizerFast"}
+{"do_lower_case": false, "unk_token": "[UNK]", "sep_token": "[SEP]", "pad_token": "[PAD]", "cls_token": "[CLS]", "mask_token": "[MASK]", "tokenize_chinese_chars": true, "strip_accents": false, "model_max_length": 512, "do_basic_tokenize": true, "tokenizer_class": "BertTokenizerFast"}