model update

config.json

@@ -1,5 +1,5 @@
 {
-  "_name_or_path": "cner_output/model/baseline_2021/bert_base_concat/best_model",
+  "_name_or_path": "bert-base-cased",
   "architectures": [
     "BertForTokenClassification"
   ],

eval/metric.json

@@ -0,0 +1 @@
+{"2021.dev": {"micro/f1": 0.6272365805168986, "micro/f1_ci": {}, "micro/recall": 0.631, "micro/precision": 0.6235177865612648, "macro/f1": 0.578141938973107, "macro/f1_ci": {}, "macro/recall": 0.577392851398218, "macro/precision": 0.5799685464596722, "per_entity_metric": {"corporation": {"f1": 0.5918367346938775, "f1_ci": {}, "precision": 0.6170212765957447, "recall": 0.5686274509803921}, "creative_work": {"f1": 0.3815789473684211, "f1_ci": {}, "precision": 0.3717948717948718, "recall": 0.3918918918918919}, "event": {"f1": 0.41947565543071164, "f1_ci": {}, "precision": 0.4117647058823529, "recall": 0.42748091603053434}, "group": {"f1": 0.6105263157894737, "f1_ci": {}, "precision": 0.5846774193548387, "recall": 0.6387665198237885}, "location": {"f1": 0.5797101449275361, "f1_ci": {}, "precision": 0.6060606060606061, "recall": 0.5555555555555556}, "person": {"f1": 0.8270944741532976, "f1_ci": {}, "precision": 0.8345323741007195, "recall": 0.8197879858657244}, "product": {"f1": 0.6367713004484306, "f1_ci": {}, "precision": 0.6339285714285714, "recall": 0.6396396396396397}}}, "2021.test": {"micro/f1": 0.6230258640421148, "micro/f1_ci": {"90": [0.6139925448708724, 0.632549139769655], "95": [0.612303125388328, 0.6336744975616968]}, "micro/recall": 0.6295097132284921, "micro/precision": 0.6166742183960127, "macro/f1": 0.5758556427048315, "macro/f1_ci": {"90": [0.5656519179998177, 0.585371041220358], "95": [0.5639180067469939, 0.586921603961901]}, "macro/recall": 0.5821234872899773, "macro/precision": 0.5715554663683273, "per_entity_metric": {"corporation": {"f1": 0.5141176470588235, "f1_ci": {"90": [0.4878216484874022, 0.5397454481031415], "95": [0.48169465214919754, 0.5451600061071579]}, "precision": 0.54625, "recall": 0.4855555555555556}, "creative_work": {"f1": 0.3886075949367089, "f1_ci": {"90": [0.3597256986658403, 0.4188124874941639], "95": [0.3535082020207142, 0.42354681306371]}, "precision": 0.36160188457008247, "recall": 0.41997264021887826}, "event": {"f1": 0.4580617122990004, "f1_ci": {"90": [0.4348966918412935, 0.4793315743183817], "95": [0.43126880173262044, 0.48449054654247115]}, "precision": 0.4384359400998336, "recall": 0.47952684258416745}, "group": {"f1": 0.5660613650594865, "f1_ci": {"90": [0.5454485939592322, 0.5872104279571561], "95": [0.5417478676490005, 0.5923902852388108]}, "precision": 0.5393794749403341, "recall": 0.5955204216073782}, "location": {"f1": 0.6264564770390679, "f1_ci": {"90": [0.597997138769671, 0.6547394206191423], "95": [0.5912249546857395, 0.6582113395040892]}, "precision": 0.6150740242261103, "recall": 0.638268156424581}, "person": {"f1": 0.8196536144578314, "f1_ci": {"90": [0.8086892670302934, 0.8320832481435211], "95": [0.8064352860749925, 0.8343931055298054]}, "precision": 0.8373076923076923, "recall": 0.8027286135693216}, "product": {"f1": 0.6580310880829014, "f1_ci": {"90": [0.6350598078981826, 0.6793531734242207], "95": [0.6308296803788606, 0.6823582405935348]}, "precision": 0.662839248434238, "recall": 0.6532921810699589}}}, "2020.test": {"micro/f1": 0.6210070384407147, "micro/f1_ci": {"90": [0.598677414852038, 0.6408161660194358], "95": [0.5946501055829765, 0.6451765433928058]}, "micro/recall": 0.5952257394914374, "micro/precision": 0.6491228070175439, "macro/f1": 0.577436139660066, "macro/f1_ci": {"90": [0.553295935898165, 0.5979250785226665], "95": [0.5491543194937442, 0.6020438981681012]}, "macro/recall": 0.549500601374034, "macro/precision": 0.6119340101835135, "per_entity_metric": {"corporation": {"f1": 0.5654761904761905, "f1_ci": {"90": [0.5016501650165016, 0.616316099488875], "95": [0.4921110450408684, 0.62540329151051]}, "precision": 0.6551724137931034, "recall": 0.4973821989528796}, "creative_work": {"f1": 0.4152046783625731, "f1_ci": {"90": [0.353622668579627, 0.4713502460914881], "95": [0.3405208523037697, 0.4846713528693937]}, "precision": 0.43558282208588955, "recall": 0.39664804469273746}, "event": {"f1": 0.450354609929078, "f1_ci": {"90": [0.40072106996250306, 0.4983622553239021], "95": [0.39129403514561506, 0.5080412036406674]}, "precision": 0.42474916387959866, "recall": 0.47924528301886793}, "group": {"f1": 0.5423143350604491, "f1_ci": {"90": [0.48928618711385696, 0.5931441022144487], "95": [0.47662690763762394, 0.6043063856034276]}, "precision": 0.585820895522388, "recall": 0.5048231511254019}, "location": {"f1": 0.605263157894737, "f1_ci": {"90": [0.5276388094586867, 0.6763224071389902], "95": [0.5114728539985327, 0.6883279763714547]}, "precision": 0.6618705035971223, "recall": 0.5575757575757576}, "person": {"f1": 0.8185801928133217, "f1_ci": {"90": [0.7927262257895654, 0.8431634280358857], "95": [0.7871878287083727, 0.8474604661726711]}, "precision": 0.8568807339449541, "recall": 0.7835570469798657}, "product": {"f1": 0.6448598130841121, "f1_ci": {"90": [0.590887445887446, 0.6986353389977458], "95": [0.5809278885848259, 0.7056143195046112]}, "precision": 0.6634615384615384, "recall": 0.6272727272727273}}}, "2021.test (span detection)": {"micro/f1": 0.7661839619941617, "micro/f1_ci": {}, "micro/recall": 0.7740256736440384, "micro/precision": 0.7584995466908432, "macro/f1": 0.7661839619941617, "macro/f1_ci": {}, "macro/recall": 0.7740256736440384, "macro/precision": 0.7584995466908432}, "2020.test (span detection)": {"micro/f1": 0.7298321602598808, "micro/f1_ci": {}, "micro/recall": 0.6995329527763363, "micro/precision": 0.7628749292586304, "macro/f1": 0.7298321602598808, "macro/f1_ci": {}, "macro/recall": 0.6995329527763363, "macro/precision": 0.7628749292586304}}

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:6988e5108f30f2610754c3b5c1d9c52d127cae9e11f7a324bf0d46026ac50d65
-size 431006385
+oid sha256:b142b70f0682eb7d488f412faaaf8342a1719c4ea74bfe17cc225f10ab189035
+size 431009137

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, "name_or_path": "cner_output/model/baseline_2021/bert_base_concat/best_model", "special_tokens_map_file": "cner_output/model/baseline_2021/bert_base_concat/best_model/special_tokens_map.json", "tokenizer_class": "BertTokenizer"}
+{"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, "name_or_path": "bert-base-cased", "tokenizer_class": "BertTokenizer"}

trainer_config.json

@@ -0,0 +1 @@
+{"data_split": "2020_2021.train", "model": "bert-base-cased", "crf": true, "max_length": 128, "epoch": 30, "batch_size": 32, "lr": 0.0001, "random_seed": 0, "gradient_accumulation_steps": 1, "weight_decay": 1e-07, "lr_warmup_step_ratio": 0.3, "max_grad_norm": 1}