model improved

config.json

@@ -8,112 +8,176 @@
   "hidden_dropout_prob": 0.1,
   "hidden_size": 1024,
   "id2label": {
-    "0": "B-SYM",
-    "1": "B-INTJ",
-    "2": "I-ADP",
-    "3": "I-X",
-    "4": "ADP",
-    "5": "PUNCT",
-    "6": "B-VERB",
-    "7": "I-VERB",
-    "8": "I-NUM",
-    "9": "VERB",
-    "10": "PRON",
-    "11": "I-DET",
-    "12": "B-ADP",
-    "13": "PROPN",
-    "14": "I-PUNCT",
-    "15": "I-CCONJ",
-    "16": "NUM",
-    "17": "I-INTJ",
-    "18": "AUX",
-    "19": "INTJ",
-    "20": "CCONJ",
-    "21": "I-PRON",
-    "22": "B-CCONJ",
-    "23": "X",
-    "24": "B-PUNCT",
-    "25": "I-SYM",
-    "26": "I-SCONJ",
-    "27": "SCONJ",
-    "28": "NOUN",
-    "29": "DET",
-    "30": "ADV",
-    "31": "PART",
-    "32": "B-PRON",
-    "33": "I-AUX",
-    "34": "B-NUM",
-    "35": "I-ADJ",
-    "36": "B-SCONJ",
-    "37": "I-PART",
-    "38": "I-NOUN",
-    "39": "I-ADV",
-    "40": "ADJ",
-    "41": "B-X",
-    "42": "B-AUX",
-    "43": "B-PROPN",
-    "44": "B-DET",
-    "45": "B-ADV",
-    "46": "I-PROPN",
-    "47": "B-NOUN",
-    "48": "SYM",
-    "49": "B-PART",
-    "50": "B-ADJ"
+    "0": "ADJ",
+    "1": "ADJ+VERB",
+    "2": "ADP",
+    "3": "ADP+ADP",
+    "4": "ADP+VERB",
+    "5": "ADV",
+    "6": "AUX",
+    "7": "B-ADJ",
+    "8": "B-ADJ+VERB",
+    "9": "B-ADP",
+    "10": "B-ADP+ADJ",
+    "11": "B-ADP+NOUN+ADP",
+    "12": "B-ADV",
+    "13": "B-AUX",
+    "14": "B-AUX+AUX",
+    "15": "B-AUX+NOUN",
+    "16": "B-CCONJ",
+    "17": "B-DET",
+    "18": "B-INTJ",
+    "19": "B-NOUN",
+    "20": "B-NOUN+ADP",
+    "21": "B-NOUN+NOUN",
+    "22": "B-NOUN+PUNCT",
+    "23": "B-NUM",
+    "24": "B-NUM+NOUN",
+    "25": "B-PART",
+    "26": "B-PRON",
+    "27": "B-PROPN",
+    "28": "B-PROPN+ADP",
+    "29": "B-PROPN+PROPN",
+    "30": "B-PUNCT",
+    "31": "B-SCONJ",
+    "32": "B-SYM",
+    "33": "B-VERB",
+    "34": "B-VERB+AUX",
+    "35": "B-VERB+SCONJ",
+    "36": "B-X",
+    "37": "CCONJ",
+    "38": "DET",
+    "39": "DET+NOUN",
+    "40": "I-ADJ",
+    "41": "I-ADJ+VERB",
+    "42": "I-ADP",
+    "43": "I-ADP+ADJ",
+    "44": "I-ADP+NOUN+ADP",
+    "45": "I-ADV",
+    "46": "I-AUX",
+    "47": "I-AUX+AUX",
+    "48": "I-AUX+NOUN",
+    "49": "I-CCONJ",
+    "50": "I-DET",
+    "51": "I-INTJ",
+    "52": "I-NOUN",
+    "53": "I-NOUN+ADP",
+    "54": "I-NOUN+NOUN",
+    "55": "I-NOUN+PUNCT",
+    "56": "I-NUM",
+    "57": "I-NUM+NOUN",
+    "58": "I-PART",
+    "59": "I-PRON",
+    "60": "I-PROPN",
+    "61": "I-PROPN+ADP",
+    "62": "I-PROPN+PROPN",
+    "63": "I-PUNCT",
+    "64": "I-SCONJ",
+    "65": "I-SYM",
+    "66": "I-VERB",
+    "67": "I-VERB+AUX",
+    "68": "I-VERB+SCONJ",
+    "69": "I-X",
+    "70": "INTJ",
+    "71": "NOUN",
+    "72": "NUM",
+    "73": "PART",
+    "74": "PRON",
+    "75": "PROPN",
+    "76": "PUNCT",
+    "77": "SCONJ",
+    "78": "SYM",
+    "79": "SYM+PUNCT",
+    "80": "VERB",
+    "81": "VERB+AUX",
+    "82": "X"
   },
   "initializer_range": 0.02,
   "intermediate_size": 4096,
   "label2id": {
-    "ADJ": 40,
-    "ADP": 4,
-    "ADV": 30,
-    "AUX": 18,
-    "B-ADJ": 50,
-    "B-ADP": 12,
-    "B-ADV": 45,
-    "B-AUX": 42,
-    "B-CCONJ": 22,
-    "B-DET": 44,
-    "B-INTJ": 1,
-    "B-NOUN": 47,
-    "B-NUM": 34,
-    "B-PART": 49,
-    "B-PRON": 32,
-    "B-PROPN": 43,
-    "B-PUNCT": 24,
-    "B-SCONJ": 36,
-    "B-SYM": 0,
-    "B-VERB": 6,
-    "B-X": 41,
-    "CCONJ": 20,
-    "DET": 29,
-    "I-ADJ": 35,
-    "I-ADP": 2,
-    "I-ADV": 39,
-    "I-AUX": 33,
-    "I-CCONJ": 15,
-    "I-DET": 11,
-    "I-INTJ": 17,
-    "I-NOUN": 38,
-    "I-NUM": 8,
-    "I-PART": 37,
-    "I-PRON": 21,
-    "I-PROPN": 46,
-    "I-PUNCT": 14,
-    "I-SCONJ": 26,
-    "I-SYM": 25,
-    "I-VERB": 7,
-    "I-X": 3,
-    "INTJ": 19,
-    "NOUN": 28,
-    "NUM": 16,
-    "PART": 31,
-    "PRON": 10,
-    "PROPN": 13,
-    "PUNCT": 5,
-    "SCONJ": 27,
-    "SYM": 48,
-    "VERB": 9,
-    "X": 23
+    "ADJ": 0,
+    "ADJ+VERB": 1,
+    "ADP": 2,
+    "ADP+ADP": 3,
+    "ADP+VERB": 4,
+    "ADV": 5,
+    "AUX": 6,
+    "B-ADJ": 7,
+    "B-ADJ+VERB": 8,
+    "B-ADP": 9,
+    "B-ADP+ADJ": 10,
+    "B-ADP+NOUN+ADP": 11,
+    "B-ADV": 12,
+    "B-AUX": 13,
+    "B-AUX+AUX": 14,
+    "B-AUX+NOUN": 15,
+    "B-CCONJ": 16,
+    "B-DET": 17,
+    "B-INTJ": 18,
+    "B-NOUN": 19,
+    "B-NOUN+ADP": 20,
+    "B-NOUN+NOUN": 21,
+    "B-NOUN+PUNCT": 22,
+    "B-NUM": 23,
+    "B-NUM+NOUN": 24,
+    "B-PART": 25,
+    "B-PRON": 26,
+    "B-PROPN": 27,
+    "B-PROPN+ADP": 28,
+    "B-PROPN+PROPN": 29,
+    "B-PUNCT": 30,
+    "B-SCONJ": 31,
+    "B-SYM": 32,
+    "B-VERB": 33,
+    "B-VERB+AUX": 34,
+    "B-VERB+SCONJ": 35,
+    "B-X": 36,
+    "CCONJ": 37,
+    "DET": 38,
+    "DET+NOUN": 39,
+    "I-ADJ": 40,
+    "I-ADJ+VERB": 41,
+    "I-ADP": 42,
+    "I-ADP+ADJ": 43,
+    "I-ADP+NOUN+ADP": 44,
+    "I-ADV": 45,
+    "I-AUX": 46,
+    "I-AUX+AUX": 47,
+    "I-AUX+NOUN": 48,
+    "I-CCONJ": 49,
+    "I-DET": 50,
+    "I-INTJ": 51,
+    "I-NOUN": 52,
+    "I-NOUN+ADP": 53,
+    "I-NOUN+NOUN": 54,
+    "I-NOUN+PUNCT": 55,
+    "I-NUM": 56,
+    "I-NUM+NOUN": 57,
+    "I-PART": 58,
+    "I-PRON": 59,
+    "I-PROPN": 60,
+    "I-PROPN+ADP": 61,
+    "I-PROPN+PROPN": 62,
+    "I-PUNCT": 63,
+    "I-SCONJ": 64,
+    "I-SYM": 65,
+    "I-VERB": 66,
+    "I-VERB+AUX": 67,
+    "I-VERB+SCONJ": 68,
+    "I-X": 69,
+    "INTJ": 70,
+    "NOUN": 71,
+    "NUM": 72,
+    "PART": 73,
+    "PRON": 74,
+    "PROPN": 75,
+    "PUNCT": 76,
+    "SCONJ": 77,
+    "SYM": 78,
+    "SYM+PUNCT": 79,
+    "VERB": 80,
+    "VERB+AUX": 81,
+    "X": 82
   },
   "layer_norm_eps": 1e-12,
   "max_position_embeddings": 512,
@@ -122,9 +186,144 @@
   "num_hidden_layers": 24,
   "pad_token_id": 0,
   "position_embedding_type": "absolute",
+  "task_specific_params": {
+    "upos_multiword": {
+      "ADJ+VERB": {
+        "\u7a0b\u306a\u304f\u3057": [
+          "\u7a0b\u306a\u304f",
+          "\u3057"
+        ]
+      },
+      "ADP+ADJ": {
+        "\u306f\u3089\u30fb\u3080\u3046\u3093\u4f5c\u54c1\u5171\u901a": [
+          "\u306f",
+          "\u3089\u30fb\u3080\u3046\u3093\u4f5c\u54c1\u5171\u901a"
+        ]
+      },
+      "ADP+ADP": {
+        "\u3068\u3082": [
+          "\u3068",
+          "\u3082"
+        ]
+      },
+      "ADP+NOUN+ADP": {
+        "\u306e\u307f\u305d\u306e": [
+          "\u306e",
+          "\u307f\u305d",
+          "\u306e"
+        ]
+      },
+      "ADP+VERB": {
+        "\u3067\u304d": [
+          "\u3067",
+          "\u304d"
+        ],
+        "\u3067\u3057": [
+          "\u3067",
+          "\u3057"
+        ],
+        "\u306f\u3057": [
+          "\u306f",
+          "\u3057"
+        ]
+      },
+      "AUX+AUX": {
+        "\u30c1\u30e3\u30c3\u30bf": [
+          "\u30c1\u30e3\u30c3",
+          "\u30bf"
+        ]
+      },
+      "AUX+NOUN": {
+        "\u306a\u304a\u5024\u6bb5": [
+          "\u306a",
+          "\u304a\u5024\u6bb5"
+        ],
+        "\u306a\u304a\u5e03\u65bd": [
+          "\u306a",
+          "\u304a\u5e03\u65bd"
+        ]
+      },
+      "DET+NOUN": {
+        "\u3053\u306e\u9803": [
+          "\u3053\u306e",
+          "\u9803"
+        ]
+      },
+      "NOUN+ADP": {
+        "\u3046\u305d\u306e": [
+          "\u3046\u305d",
+          "\u306e"
+        ]
+      },
+      "NOUN+NOUN": {
+        "\u9ce5\u53d6\u770c\u6559\u80b2\u59d4\u54e1\u4f1a\u793e\u4f1a\u6559\u80b2\u59d4\u54e1": [
+          "\u9ce5\u53d6\u770c\u6559\u80b2\u59d4\u54e1\u4f1a",
+          "\u793e\u4f1a\u6559\u80b2\u59d4\u54e1"
+        ]
+      },
+      "NOUN+PUNCT": {
+        "\u5b87\u5b99\u4ea4\u97ff\u8a69-\u300f": [
+          "\u5b87\u5b99\u4ea4\u97ff\u8a69-",
+          "\u300f"
+        ],
+        "\u7259\u72fc-GARO-\u300f": [
+          "\u7259\u72fc-GARO-",
+          "\u300f"
+        ]
+      },
+      "NUM+NOUN": {
+        "3.0LV6\u30a8\u30f3\u30b8\u30f3": [
+          "3.0L",
+          "V6\u30a8\u30f3\u30b8\u30f3"
+        ],
+        "4.7LV8\u30a8\u30f3\u30b8\u30f3": [
+          "4.7L",
+          "V8\u30a8\u30f3\u30b8\u30f3"
+        ]
+      },
+      "PROPN+ADP": {
+        "\u3061\u307b\u306e": [
+          "\u3061\u307b",
+          "\u306e"
+        ],
+        "\u3088\u3090\u3053\u306e": [
+          "\u3088\u3090\u3053",
+          "\u306e"
+        ]
+      },
+      "PROPN+PROPN": {
+        "\u5e83\u5cf6\u5e02\u4e2d\u533a": [
+          "\u5e83\u5cf6\u5e02",
+          "\u4e2d\u533a"
+        ]
+      },
+      "SYM+PUNCT": {
+        "-)": [
+          "-",
+          ")"
+        ]
+      },
+      "VERB+AUX": {
+        "\u305f\u3063\u305f": [
+          "\u305f\u3063",
+          "\u305f"
+        ],
+        "\u5c0f\u5c4b\u639b\u3051\u3055\u305b": [
+          "\u5c0f\u5c4b\u639b\u3051\u3055",
+          "\u305b"
+        ]
+      },
+      "VERB+SCONJ": {
+        "\u8cb8\u5207\u3063\u3066": [
+          "\u8cb8\u5207\u3063",
+          "\u3066"
+        ]
+      }
+    }
+  },
   "tokenizer_class": "BertJapaneseTokenizer",
   "torch_dtype": "float32",
-  "transformers_version": "4.11.3",
+  "transformers_version": "4.19.2",
   "type_vocab_size": 2,
   "use_cache": true,
   "vocab_size": 32768

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c26b02e8926191ac7f7fb3340d3d286e08fe213b8b34a0dafe120ad0b2f00bec
-size 1345944210
+oid sha256:215468db9aeba6a4f9bfc28f4e7ea62924a41734f78c6454114ac3edbd0826b4
+size 1346049874

supar.model

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:b2e1383c4cbb6c4a5f2a2a09cb8a89e726415291e030194772840dbcc68f68d9
-size 1399761610
+oid sha256:638e2ba0224ebebca9801f7a47e3327cae29e0a2889e40bc5eb61642d2efb903
+size 1398774350

tokenizer_config.json

@@ -1 +1 @@
-{"unk_token": "[UNK]", "sep_token": "[SEP]", "pad_token": "[PAD]", "cls_token": "[CLS]", "mask_token": "[MASK]", "do_lower_case": false, "do_word_tokenize": true, "do_subword_tokenize": true, "word_tokenizer_type": "mecab", "subword_tokenizer_type": "wordpiece", "never_split": null, "mecab_kwargs": {"mecab_dic": "unidic_lite"}, "model_max_length": 512, "tokenizer_class": "BertJapaneseTokenizer"}
+{"unk_token": "[UNK]", "sep_token": "[SEP]", "pad_token": "[PAD]", "cls_token": "[CLS]", "mask_token": "[MASK]", "do_lower_case": false, "do_word_tokenize": true, "do_subword_tokenize": true, "word_tokenizer_type": "mecab", "subword_tokenizer_type": "wordpiece", "never_split": null, "mecab_kwargs": {"mecab_dic": "unidic_lite"}, "model_max_length": 512, "tokenizer_class": "BertJapaneseTokenizerFast"}