model improved

README.md

@@ -19,7 +19,7 @@ widget:
 
 ## Model Description
 
-This is a BERT model pre-trained on Japanese Wikipedia texts for POS-tagging and dependency-parsing, derived from [bert-large-japanese-char-extended](https://huggingface.co/KoichiYasuoka/bert-large-japanese-char-extended). Every long-unit-word is tagged by [UPOS](https://universaldependencies.org/u/pos/) (Universal Part-Of-Speech).
+This is a BERT model pre-trained on Japanese Wikipedia texts for POS-tagging and dependency-parsing, derived from [bert-large-japanese-char-extended](https://huggingface.co/KoichiYasuoka/bert-large-japanese-char-extended). Every long-unit-word is tagged by [UPOS](https://universaldependencies.org/u/pos/) (Universal Part-Of-Speech) and [FEATS](https://universaldependencies.org/u/feat/).
 
 ## How to Use
 

config.json

@@ -13,51 +13,56 @@
     "1": "ADP",
     "2": "ADV",
     "3": "AUX",
-    "4": "B-ADJ",
-    "5": "B-ADP",
-    "6": "B-ADV",
-    "7": "B-AUX",
-    "8": "B-CCONJ",
-    "9": "B-DET",
-    "10": "B-INTJ",
-    "11": "B-NOUN",
-    "12": "B-NUM",
-    "13": "B-PART",
-    "14": "B-PRON",
-    "15": "B-PROPN",
-    "16": "B-PUNCT",
-    "17": "B-SCONJ",
-    "18": "B-SYM",
-    "19": "B-VERB",
-    "20": "B-X",
-    "21": "CCONJ",
-    "22": "I-ADJ",
-    "23": "I-ADP",
-    "24": "I-ADV",
-    "25": "I-AUX",
-    "26": "I-CCONJ",
-    "27": "I-DET",
-    "28": "I-INTJ",
-    "29": "I-NOUN",
-    "30": "I-NUM",
-    "31": "I-PART",
-    "32": "I-PRON",
-    "33": "I-PROPN",
-    "34": "I-PUNCT",
-    "35": "I-SCONJ",
-    "36": "I-SYM",
-    "37": "I-VERB",
-    "38": "I-X",
-    "39": "NOUN",
-    "40": "NUM",
-    "41": "PART",
-    "42": "PRON",
-    "43": "PROPN",
-    "44": "PUNCT",
-    "45": "SCONJ",
-    "46": "SYM",
-    "47": "VERB",
-    "48": "X"
+    "4": "AUX|Polarity=Neg",
+    "5": "B-ADJ",
+    "6": "B-ADP",
+    "7": "B-ADV",
+    "8": "B-AUX",
+    "9": "B-AUX|Polarity=Neg",
+    "10": "B-CCONJ",
+    "11": "B-DET",
+    "12": "B-INTJ",
+    "13": "B-NOUN",
+    "14": "B-NOUN|Polarity=Neg",
+    "15": "B-NUM",
+    "16": "B-PART",
+    "17": "B-PRON",
+    "18": "B-PROPN",
+    "19": "B-PUNCT",
+    "20": "B-SCONJ",
+    "21": "B-SYM",
+    "22": "B-VERB",
+    "23": "B-X",
+    "24": "CCONJ",
+    "25": "I-ADJ",
+    "26": "I-ADP",
+    "27": "I-ADV",
+    "28": "I-AUX",
+    "29": "I-AUX|Polarity=Neg",
+    "30": "I-CCONJ",
+    "31": "I-DET",
+    "32": "I-INTJ",
+    "33": "I-NOUN",
+    "34": "I-NOUN|Polarity=Neg",
+    "35": "I-NUM",
+    "36": "I-PART",
+    "37": "I-PRON",
+    "38": "I-PROPN",
+    "39": "I-PUNCT",
+    "40": "I-SCONJ",
+    "41": "I-SYM",
+    "42": "I-VERB",
+    "43": "I-X",
+    "44": "NOUN",
+    "45": "NUM",
+    "46": "PART",
+    "47": "PRON",
+    "48": "PROPN",
+    "49": "PUNCT",
+    "50": "SCONJ",
+    "51": "SYM",
+    "52": "VERB",
+    "53": "X"
   },
   "initializer_range": 0.02,
   "intermediate_size": 4096,
@@ -66,51 +71,56 @@
     "ADP": 1,
     "ADV": 2,
     "AUX": 3,
-    "B-ADJ": 4,
-    "B-ADP": 5,
-    "B-ADV": 6,
-    "B-AUX": 7,
-    "B-CCONJ": 8,
-    "B-DET": 9,
-    "B-INTJ": 10,
-    "B-NOUN": 11,
-    "B-NUM": 12,
-    "B-PART": 13,
-    "B-PRON": 14,
-    "B-PROPN": 15,
-    "B-PUNCT": 16,
-    "B-SCONJ": 17,
-    "B-SYM": 18,
-    "B-VERB": 19,
-    "B-X": 20,
-    "CCONJ": 21,
-    "I-ADJ": 22,
-    "I-ADP": 23,
-    "I-ADV": 24,
-    "I-AUX": 25,
-    "I-CCONJ": 26,
-    "I-DET": 27,
-    "I-INTJ": 28,
-    "I-NOUN": 29,
-    "I-NUM": 30,
-    "I-PART": 31,
-    "I-PRON": 32,
-    "I-PROPN": 33,
-    "I-PUNCT": 34,
-    "I-SCONJ": 35,
-    "I-SYM": 36,
-    "I-VERB": 37,
-    "I-X": 38,
-    "NOUN": 39,
-    "NUM": 40,
-    "PART": 41,
-    "PRON": 42,
-    "PROPN": 43,
-    "PUNCT": 44,
-    "SCONJ": 45,
-    "SYM": 46,
-    "VERB": 47,
-    "X": 48
+    "AUX|Polarity=Neg": 4,
+    "B-ADJ": 5,
+    "B-ADP": 6,
+    "B-ADV": 7,
+    "B-AUX": 8,
+    "B-AUX|Polarity=Neg": 9,
+    "B-CCONJ": 10,
+    "B-DET": 11,
+    "B-INTJ": 12,
+    "B-NOUN": 13,
+    "B-NOUN|Polarity=Neg": 14,
+    "B-NUM": 15,
+    "B-PART": 16,
+    "B-PRON": 17,
+    "B-PROPN": 18,
+    "B-PUNCT": 19,
+    "B-SCONJ": 20,
+    "B-SYM": 21,
+    "B-VERB": 22,
+    "B-X": 23,
+    "CCONJ": 24,
+    "I-ADJ": 25,
+    "I-ADP": 26,
+    "I-ADV": 27,
+    "I-AUX": 28,
+    "I-AUX|Polarity=Neg": 29,
+    "I-CCONJ": 30,
+    "I-DET": 31,
+    "I-INTJ": 32,
+    "I-NOUN": 33,
+    "I-NOUN|Polarity=Neg": 34,
+    "I-NUM": 35,
+    "I-PART": 36,
+    "I-PRON": 37,
+    "I-PROPN": 38,
+    "I-PUNCT": 39,
+    "I-SCONJ": 40,
+    "I-SYM": 41,
+    "I-VERB": 42,
+    "I-X": 43,
+    "NOUN": 44,
+    "NUM": 45,
+    "PART": 46,
+    "PRON": 47,
+    "PROPN": 48,
+    "PUNCT": 49,
+    "SCONJ": 50,
+    "SYM": 51,
+    "VERB": 52,
+    "X": 53
   },
   "layer_norm_eps": 1e-12,
   "max_position_embeddings": 512,
@@ -119,9 +129,9 @@
   "num_hidden_layers": 24,
   "pad_token_id": 0,
   "position_embedding_type": "absolute",
-  "tokenizer_class": "BertTokenizer",
+  "tokenizer_class": "BertTokenizerFast",
   "torch_dtype": "float32",
-  "transformers_version": "4.19.2",
+  "transformers_version": "4.19.4",
   "type_vocab_size": 2,
   "use_cache": true,
   "vocab_size": 6291

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d374996c4a74c84205d41521dd47f0510dc723fcd5b378bb68ee0fde2557a144
-size 1237455217
+oid sha256:5df3afc1ee26ab45efdfc78faae68535c906366599f9d473286647818bcf97fd
+size 1237475697

supar.model

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:13a3eed7e15f59a2a3c2020a506f76bbd478d9ccbe0ea534214960f7d17350ed
+oid sha256:18d822234fa93a516495b8145b7848abd64b59fc3dedfc2819c7116469ea8d4b
 size 1289789669