init

config.json

@@ -0,0 +1,42 @@
+{
+  "_name_or_path": "xlm-roberta-large",
+  "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": 1024,
+  "id2label": {
+    "0": "O",
+    "1": "B-location",
+    "2": "B-person",
+    "3": "I-person",
+    "4": "B-organization",
+    "5": "I-location",
+    "6": "I-organization"
+  },
+  "initializer_range": 0.02,
+  "intermediate_size": 4096,
+  "label2id": {
+    "B-location": 1,
+    "B-organization": 4,
+    "B-person": 2,
+    "I-location": 5,
+    "I-organization": 6,
+    "I-person": 3,
+    "O": 0
+  },
+  "layer_norm_eps": 1e-05,
+  "max_position_embeddings": 514,
+  "model_type": "xlm-roberta",
+  "num_attention_heads": 16,
+  "num_hidden_layers": 24,
+  "output_past": true,
+  "pad_token_id": 1,
+  "type_vocab_size": 1,
+  "vocab_size": 250002
+}

parameter.json

@@ -0,0 +1 @@
+{"dataset": ["panx_dataset/ru"], "transformers_model": "xlm-roberta-large", "random_seed": 1234, "lr": 1e-05, "total_step": 5000, "warmup_step": 700, "weight_decay": 1e-07, "batch_size": 32, "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 sha256:95701726e524d85cf5db68f90d747b0537c4e6c4b594e8a911f85598e27bdcbc
+size 2235557289

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_panx_dataset-ar.json

@@ -0,0 +1 @@
+{"valid": {"f1": 55.22605332113297, "recall": 58.92429218070472, "precision": 51.96462116468378, "summary": "              precision    recall  f1-score   support\n\n    location       0.56      0.56      0.56      3856\norganization       0.37      0.58      0.45      3596\n      person       0.73      0.63      0.68      3815\n\n   micro avg       0.52      0.59      0.55     11267\n   macro avg       0.55      0.59      0.56     11267\nweighted avg       0.56      0.59      0.56     11267\n"}, "test": {"f1": 54.980643549931315, "recall": 58.655297983835155, "precision": 51.73926668755876, "summary": "              precision    recall  f1-score   support\n\n    location       0.55      0.54      0.55      3780\norganization       0.38      0.60      0.46      3629\n      person       0.72      0.62      0.66      3850\n\n   micro avg       0.52      0.59      0.55     11259\n   macro avg       0.55      0.59      0.56     11259\nweighted avg       0.55      0.59      0.56     11259\n"}}

test_panx_dataset-en.json

@@ -0,0 +1 @@
+{"valid": {"f1": 60.53327241278027, "recall": 65.7031526744598, "precision": 56.11763282100932, "summary": "              precision    recall  f1-score   support\n\n    location       0.48      0.57      0.52      4803\norganization       0.49      0.70      0.58      4677\n      person       0.78      0.70      0.74      4635\n\n   micro avg       0.56      0.66      0.61     14115\n   macro avg       0.58      0.66      0.61     14115\nweighted avg       0.58      0.66      0.61     14115\n"}, "test": {"f1": 60.492558586628654, "recall": 65.67583129408378, "precision": 56.06758832565284, "summary": "              precision    recall  f1-score   support\n\n    location       0.48      0.56      0.52      4633\norganization       0.51      0.71      0.59      4744\n      person       0.76      0.70      0.73      4517\n\n   micro avg       0.56      0.66      0.60     13894\n   macro avg       0.58      0.66      0.61     13894\nweighted avg       0.58      0.66      0.61     13894\n"}}

test_panx_dataset-es.json

@@ -0,0 +1 @@
+{"valid": {"f1": 76.24733138293666, "recall": 78.59005704971476, "precision": 74.04023341523342, "summary": "              precision    recall  f1-score   support\n\n    location       0.83      0.78      0.81      4567\norganization       0.58      0.85      0.69      3737\n      person       0.87      0.73      0.80      3966\n\n   micro avg       0.74      0.79      0.76     12270\n   macro avg       0.76      0.79      0.77     12270\nweighted avg       0.77      0.79      0.77     12270\n"}, "test": {"f1": 76.88602897635309, "recall": 79.42903752039152, "precision": 74.5008033050264, "summary": "              precision    recall  f1-score   support\n\n    location       0.84      0.79      0.82      4725\norganization       0.59      0.86      0.70      3576\n      person       0.86      0.74      0.80      3959\n\n   micro avg       0.75      0.79      0.77     12260\n   macro avg       0.76      0.80      0.77     12260\nweighted avg       0.78      0.79      0.78     12260\n"}}

test_panx_dataset-ja.json

@@ -0,0 +1 @@
+{"valid": {"f1": 53.384072674002546, "recall": 61.58388531691303, "precision": 47.11126961483595, "summary": "              precision    recall  f1-score   support\n\n    location       0.54      0.69      0.61     19595\norganization       0.34      0.60      0.43     16328\n      person       0.63      0.55      0.59     15558\n\n   micro avg       0.47      0.62      0.53     51481\n   macro avg       0.50      0.61      0.54     51481\nweighted avg       0.50      0.62      0.54     51481\n"}, "test": {"f1": 53.384072674002546, "recall": 61.58388531691303, "precision": 47.11126961483595, "summary": "              precision    recall  f1-score   support\n\n    location       0.54      0.69      0.61     19595\norganization       0.34      0.60      0.43     16328\n      person       0.63      0.55      0.59     15558\n\n   micro avg       0.47      0.62      0.53     51481\n   macro avg       0.50      0.61      0.54     51481\nweighted avg       0.50      0.62      0.54     51481\n"}}

test_panx_dataset-ko.json

@@ -0,0 +1 @@
+{"valid": {"f1": 69.14568906101283, "recall": 72.36381122484028, "precision": 66.20160884866768, "summary": "              precision    recall  f1-score   support\n\n    location       0.79      0.80      0.79      5997\norganization       0.48      0.67      0.56      4459\n      person       0.76      0.67      0.71      4101\n\n   micro avg       0.66      0.72      0.69     14557\n   macro avg       0.68      0.71      0.69     14557\nweighted avg       0.69      0.72      0.70     14557\n"}, "test": {"f1": 68.12990366471348, "recall": 71.34438050336269, "precision": 65.19260010136847, "summary": "              precision    recall  f1-score   support\n\n    location       0.78      0.79      0.78      5855\norganization       0.48      0.67      0.56      4319\n      person       0.74      0.65      0.69      4249\n\n   micro avg       0.65      0.71      0.68     14423\n   macro avg       0.66      0.70      0.68     14423\nweighted avg       0.68      0.71      0.69     14423\n"}}

test_panx_dataset-ru.json

@@ -0,0 +1 @@
+{"valid": {"f1": 90.48385798244908, "recall": 91.1221015080266, "precision": 89.85449312440038, "summary": "              precision    recall  f1-score   support\n\n    location       0.91      0.92      0.91      4852\norganization       0.84      0.85      0.85      3892\n      person       0.95      0.96      0.96      3590\n\n   micro avg       0.90      0.91      0.90     12334\n   macro avg       0.90      0.91      0.91     12334\nweighted avg       0.90      0.91      0.90     12334\n"}, "test": {"f1": 90.07135575942915, "recall": 90.70378582573704, "precision": 89.44768383543894, "summary": "              precision    recall  f1-score   support\n\n    location       0.90      0.92      0.91      4560\norganization       0.84      0.85      0.84      4074\n      person       0.94      0.96      0.95      3543\n\n   micro avg       0.89      0.91      0.90     12177\n   macro avg       0.90      0.91      0.90     12177\nweighted avg       0.89      0.91      0.90     12177\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-large"}