init

config.json

@@ -0,0 +1,102 @@
+{
+  "_name_or_path": "xlm-roberta-base",
+  "architectures": [
+    "XLMRobertaForTokenClassification"
+  ],
+  "attention_probs_dropout_prob": 0.1,
+  "bos_token_id": 0,
+  "eos_token_id": 2,
+  "gradient_checkpointing": false,
+  "hidden_act": "gelu",
+  "hidden_dropout_prob": 0.1,
+  "hidden_size": 768,
+  "id2label": {
+    "0": "O",
+    "1": "B-cardinal number",
+    "2": "B-date",
+    "3": "I-date",
+    "4": "B-person",
+    "5": "I-person",
+    "6": "B-group",
+    "7": "B-geopolitical area",
+    "8": "I-geopolitical area",
+    "9": "B-law",
+    "10": "I-law",
+    "11": "B-organization",
+    "12": "I-organization",
+    "13": "B-percent",
+    "14": "I-percent",
+    "15": "B-ordinal number",
+    "16": "B-money",
+    "17": "I-money",
+    "18": "B-work of art",
+    "19": "I-work of art",
+    "20": "B-facility",
+    "21": "B-time",
+    "22": "I-cardinal number",
+    "23": "B-location",
+    "24": "B-quantity",
+    "25": "I-quantity",
+    "26": "I-group",
+    "27": "I-location",
+    "28": "B-product",
+    "29": "I-time",
+    "30": "B-event",
+    "31": "I-event",
+    "32": "I-facility",
+    "33": "B-language",
+    "34": "I-product",
+    "35": "I-ordinal number",
+    "36": "I-language"
+  },
+  "initializer_range": 0.02,
+  "intermediate_size": 3072,
+  "label2id": {
+    "B-cardinal number": 1,
+    "B-date": 2,
+    "B-event": 30,
+    "B-facility": 20,
+    "B-geopolitical area": 7,
+    "B-group": 6,
+    "B-language": 33,
+    "B-law": 9,
+    "B-location": 23,
+    "B-money": 16,
+    "B-ordinal number": 15,
+    "B-organization": 11,
+    "B-percent": 13,
+    "B-person": 4,
+    "B-product": 28,
+    "B-quantity": 24,
+    "B-time": 21,
+    "B-work of art": 18,
+    "I-cardinal number": 22,
+    "I-date": 3,
+    "I-event": 31,
+    "I-facility": 32,
+    "I-geopolitical area": 8,
+    "I-group": 26,
+    "I-language": 36,
+    "I-law": 10,
+    "I-location": 27,
+    "I-money": 17,
+    "I-ordinal number": 35,
+    "I-organization": 12,
+    "I-percent": 14,
+    "I-person": 5,
+    "I-product": 34,
+    "I-quantity": 25,
+    "I-time": 29,
+    "I-work of art": 19,
+    "O": 0
+  },
+  "layer_norm_eps": 1e-05,
+  "max_position_embeddings": 514,
+  "model_type": "xlm-roberta",
+  "num_attention_heads": 12,
+  "num_hidden_layers": 12,
+  "output_past": true,
+  "pad_token_id": 1,
+  "type_vocab_size": 1,
+  "vocab_size": 250002
+}

parameter.json

@@ -0,0 +1 @@
+{"dataset": ["ontonotes5"], "transformers_model": "xlm-roberta-base", "random_seed": 1234, "lr": 1e-05, "total_step": 13000, "warmup_step": 700, "weight_decay": 1e-07, "batch_size": 16, "max_seq_length": 128, "fp16": false, "max_grad_norm": 1.0, "lower_case": false}

pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:13c17f9d9b4742ac5aa7f0e56886c3b4fef04737f67efb63730616a17cf39682
+size 1110011650

sentencepiece.bpe.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cfc8146abe2a0488e9e2a0c56de7952f7c11ab059eca145a0a727afce0db2865
+size 5069051

special_tokens_map.json

@@ -0,0 +1 @@
+{"bos_token": "<s>", "eos_token": "</s>", "unk_token": "<unk>", "sep_token": "</s>", "pad_token": "<pad>", "cls_token": "<s>", "mask_token": "<mask>"}

test_bc5cdr_span.json

@@ -0,0 +1 @@
+{"valid": {"f1": 0.0, "recall": 0.0, "precision": 0.0, "summary": ""}, "test": {"f1": 0.0, "recall": 0.0, "precision": 0.0, "summary": ""}}

test_bionlp2004_span.json

@@ -0,0 +1 @@
+{"valid": {"f1": 0.0, "recall": 0.0, "precision": 0.0, "summary": ""}}

test_conll2003_span.json

@@ -0,0 +1 @@
+{"valid": {"f1": 64.09021794221997, "recall": 50.54967019788127, "precision": 87.53894080996885, "summary": "              precision    recall  f1-score   support\n\n      entity       0.88      0.51      0.64      5003\n\n   micro avg       0.88      0.51      0.64      5003\n   macro avg       0.88      0.51      0.64      5003\nweighted avg       0.88      0.51      0.64      5003\n"}, "test": {"f1": 62.21701795472286, "recall": 48.48915027377814, "precision": 86.78765880217786, "summary": "              precision    recall  f1-score   support\n\n      entity       0.87      0.48      0.62      4931\n\n   micro avg       0.87      0.48      0.62      4931\n   macro avg       0.87      0.48      0.62      4931\nweighted avg       0.87      0.48      0.62      4931\n"}}

test_fin_span.json

@@ -0,0 +1 @@
+{"valid": {"f1": 31.753554502369667, "recall": 26.58730158730159, "precision": 39.411764705882355, "summary": "              precision    recall  f1-score   support\n\n      entity       0.39      0.27      0.32       252\n\n   micro avg       0.39      0.27      0.32       252\n   macro avg       0.39      0.27      0.32       252\nweighted avg       0.39      0.27      0.32       252\n"}}

test_ontonotes5.json

@@ -0,0 +1 @@
+{"valid": {"f1": 86.98359051896652, "recall": 88.57893306765692, "precision": 85.4446968373231, "summary": "                   precision    recall  f1-score   support\n\n  cardinal number       0.82      0.88      0.85       937\n             date       0.83      0.87      0.85      1507\n            event       0.63      0.55      0.59       143\n         facility       0.54      0.64      0.59       115\ngeopolitical area       0.94      0.93      0.94      2262\n            group       0.90      0.93      0.91       847\n         language       0.86      0.73      0.79        33\n              law       0.46      0.70      0.55        40\n         location       0.67      0.74      0.70       204\n            money       0.90      0.93      0.92       274\n   ordinal number       0.83      0.86      0.85       232\n     organization       0.86      0.86      0.86      1728\n          percent       0.89      0.89      0.89       177\n           person       0.91      0.96      0.93      2014\n          product       0.55      0.67      0.60        72\n         quantity       0.79      0.81      0.80       100\n             time       0.68      0.79      0.73       214\n      work of art       0.42      0.54      0.48       142\n\n        micro avg       0.85      0.89      0.87     11041\n        macro avg       0.75      0.79      0.77     11041\n     weighted avg       0.86      0.89      0.87     11041\n"}, "test": {"f1": 89.0153671030165, "recall": 90.35641276330992, "precision": 87.71354616048318, "summary": "                   precision    recall  f1-score   support\n\n  cardinal number       0.85      0.88      0.86       934\n             date       0.84      0.89      0.86      1601\n            event       0.61      0.65      0.63        63\n         facility       0.76      0.74      0.75       135\ngeopolitical area       0.96      0.96      0.96      2240\n            group       0.89      0.94      0.92       841\n         language       0.75      0.55      0.63        22\n              law       0.63      0.60      0.62        40\n         location       0.70      0.80      0.74       179\n            money       0.85      0.90      0.87       314\n   ordinal number       0.81      0.92      0.86       195\n     organization       0.89      0.89      0.89      1792\n          percent       0.89      0.92      0.90       348\n           person       0.93      0.96      0.95      1988\n          product       0.65      0.72      0.69        76\n         quantity       0.77      0.81      0.79       105\n             time       0.60      0.66      0.63       212\n      work of art       0.60      0.60      0.60       166\n\n        micro avg       0.88      0.90      0.89     11251\n        macro avg       0.78      0.80      0.79     11251\n     weighted avg       0.88      0.90      0.89     11251\n"}}

test_ontonotes5_span.json

@@ -0,0 +1 @@
+{"valid": {"f1": 91.05923313841896, "recall": 92.38293632823114, "precision": 89.77292730153142, "summary": "              precision    recall  f1-score   support\n\n      entity       0.90      0.92      0.91     11041\n\n   micro avg       0.90      0.92      0.91     11041\n   macro avg       0.90      0.92      0.91     11041\nweighted avg       0.90      0.92      0.91     11041\n"}, "test": {"f1": 91.83754116355654, "recall": 92.95173762332237, "precision": 90.7497396737244, "summary": "              precision    recall  f1-score   support\n\n      entity       0.91      0.93      0.92     11251\n\n   micro avg       0.91      0.93      0.92     11251\n   macro avg       0.91      0.93      0.92     11251\nweighted avg       0.91      0.93      0.92     11251\n"}}

test_panx_dataset-en_span.json

@@ -0,0 +1 @@
+{"valid": {"f1": 44.32529043789097, "recall": 35.13992206872122, "precision": 60.01209921355112, "summary": "              precision    recall  f1-score   support\n\n      entity       0.60      0.35      0.44     14115\n\n   micro avg       0.60      0.35      0.44     14115\n   macro avg       0.60      0.35      0.44     14115\nweighted avg       0.60      0.35      0.44     14115\n"}, "test": {"f1": 44.73243010557796, "recall": 35.526126385490144, "precision": 60.379204892966364, "summary": "              precision    recall  f1-score   support\n\n      entity       0.60      0.36      0.45     13894\n\n   micro avg       0.60      0.36      0.45     13894\n   macro avg       0.60      0.36      0.45     13894\nweighted avg       0.60      0.36      0.45     13894\n"}}

test_wnut2017_span.json

@@ -0,0 +1 @@
+{"valid": {"f1": 62.40928882438316, "recall": 53.61596009975062, "precision": 74.65277777777779, "summary": "              precision    recall  f1-score   support\n\n      entity       0.75      0.54      0.62       802\n\n   micro avg       0.75      0.54      0.62       802\n   macro avg       0.75      0.54      0.62       802\nweighted avg       0.75      0.54      0.62       802\n"}, "test": {"f1": 51.70387779083432, "recall": 43.52126607319486, "precision": 63.67583212735166, "summary": "              precision    recall  f1-score   support\n\n      entity       0.64      0.44      0.52      1011\n\n   micro avg       0.64      0.44      0.52      1011\n   macro avg       0.64      0.44      0.52      1011\nweighted avg       0.64      0.44      0.52      1011\n"}}

tokenizer_config.json

@@ -0,0 +1 @@
+{"bos_token": "<s>", "eos_token": "</s>", "sep_token": "</s>", "cls_token": "<s>", "unk_token": "<unk>", "pad_token": "<pad>", "mask_token": "<mask>", "model_max_length": 512, "name_or_path": "xlm-roberta-base"}