model update

config.json

@@ -1,5 +1,5 @@
 {
-  "_name_or_path": "cner_output/model/baseline_2021/bertweet_large_concat/best_model",
+  "_name_or_path": "vinai/bertweet-large",
   "architectures": [
     "RobertaForTokenClassification"
   ],

eval/metric.json

@@ -0,0 +1 @@
+{"2021.dev": {"micro/f1": 0.6507224713502741, "micro/f1_ci": {}, "micro/recall": 0.653, "micro/precision": 0.6484607745779544, "macro/f1": 0.6103997713771362, "macro/f1_ci": {}, "macro/recall": 0.6181336282143709, "macro/precision": 0.606567588228242, "per_entity_metric": {"corporation": {"f1": 0.5841584158415842, "f1_ci": {}, "precision": 0.59, "recall": 0.5784313725490197}, "creative_work": {"f1": 0.4733727810650888, "f1_ci": {}, "precision": 0.42105263157894735, "recall": 0.5405405405405406}, "event": {"f1": 0.4897959183673469, "f1_ci": {}, "precision": 0.5263157894736842, "recall": 0.4580152671755725}, "group": {"f1": 0.6057906458797327, "f1_ci": {}, "precision": 0.6126126126126126, "recall": 0.5991189427312775}, "location": {"f1": 0.6233766233766234, "f1_ci": {}, "precision": 0.5853658536585366, "recall": 0.6666666666666666}, "person": {"f1": 0.8327526132404182, "f1_ci": {}, "precision": 0.8213058419243986, "recall": 0.8445229681978799}, "product": {"f1": 0.663551401869159, "f1_ci": {}, "precision": 0.6893203883495146, "recall": 0.6396396396396397}}}, "2021.test": {"micro/f1": 0.6646206308610401, "micro/f1_ci": {"90": [0.6554125820888649, 0.6736489128168938], "95": [0.6533077908395879, 0.675252368755536]}, "micro/recall": 0.6761100832562442, "micro/precision": 0.653515144741254, "macro/f1": 0.6187282305429461, "macro/f1_ci": {"90": [0.6086007046046046, 0.6274752439680484], "95": [0.6064349641468447, 0.6292100251466678]}, "macro/recall": 0.6359356515638321, "macro/precision": 0.6069581336386037, "per_entity_metric": {"corporation": {"f1": 0.545042492917847, "f1_ci": {"90": [0.5210565034619189, 0.570455308114704], "95": [0.5163483236977213, 0.5768635272388254]}, "precision": 0.5560693641618497, "recall": 0.5344444444444445}, "creative_work": {"f1": 0.47362250879249707, "f1_ci": {"90": [0.44319138030931055, 0.5049788934117292], "95": [0.4364029194303145, 0.5107778906673237]}, "precision": 0.41435897435897434, "recall": 0.5526675786593708}, "event": {"f1": 0.4915336236090953, "f1_ci": {"90": [0.4684507714181196, 0.5143994098920723], "95": [0.46447741808246684, 0.5187658305407751]}, "precision": 0.5247933884297521, "recall": 0.462238398544131}, "group": {"f1": 0.623768877216021, "f1_ci": {"90": [0.6036063111513918, 0.6459784481752944], "95": [0.5996059953988204, 0.6488495817907584]}, "precision": 0.6217277486910995, "recall": 0.6258234519104084}, "location": {"f1": 0.6754716981132076, "f1_ci": {"90": [0.6502264241541174, 0.7009141138003405], "95": [0.643485938839921, 0.7058891849422697]}, "precision": 0.61441647597254, "recall": 0.75}, "person": {"f1": 0.8414922656960875, "f1_ci": {"90": [0.8306318756851475, 0.8517340317116006], "95": [0.8291240468904854, 0.8531652713056682]}, "precision": 0.8307581746316924, "recall": 0.8525073746312685}, "product": {"f1": 0.6801661474558671, "f1_ci": {"90": [0.6576924955753543, 0.7017577078587071], "95": [0.6537204296451585, 0.7052648077874939]}, "precision": 0.6865828092243187, "recall": 0.6738683127572016}}}, "2020.test": {"micro/f1": 0.6675704989154012, "micro/f1_ci": {"90": [0.6478782487489827, 0.686330134406657], "95": [0.643485580441, 0.689784971727218]}, "micro/recall": 0.6388168137000519, "micro/precision": 0.6990346394094265, "macro/f1": 0.6307734401161805, "macro/f1_ci": {"90": [0.6086331719773058, 0.650352710597749], "95": [0.6044641232125019, 0.6551010145551667]}, "macro/recall": 0.6085863177550797, "macro/precision": 0.6616549497337102, "per_entity_metric": {"corporation": {"f1": 0.5888594164456233, "f1_ci": {"90": [0.5313343849856409, 0.6377922295678834], "95": [0.5233514118457299, 0.6485181964142359]}, "precision": 0.5967741935483871, "recall": 0.581151832460733}, "creative_work": {"f1": 0.5524296675191815, "f1_ci": {"90": [0.49640524689644966, 0.6021000260333806], "95": [0.48969679633867275, 0.610844280240832]}, "precision": 0.5094339622641509, "recall": 0.6033519553072626}, "event": {"f1": 0.4888888888888889, "f1_ci": {"90": [0.43619555287353434, 0.5418402724977951], "95": [0.42727297146820076, 0.5512992831541219]}, "precision": 0.5260869565217391, "recall": 0.45660377358490567}, "group": {"f1": 0.5985130111524165, "f1_ci": {"90": [0.5485224195478304, 0.650063025210084], "95": [0.534749739311783, 0.6577725388960302]}, "precision": 0.7092511013215859, "recall": 0.5176848874598071}, "location": {"f1": 0.6666666666666666, "f1_ci": {"90": [0.598896163716523, 0.7267257383966245], "95": [0.5836553829874008, 0.744198416625433]}, "precision": 0.6607142857142857, "recall": 0.6727272727272727}, "person": {"f1": 0.8349178910976663, "f1_ci": {"90": [0.8089173822232536, 0.8576861436442418], "95": [0.8028489054150965, 0.8630879876393839]}, "precision": 0.8609625668449198, "recall": 0.8104026845637584}, "product": {"f1": 0.6851385390428212, "f1_ci": {"90": [0.6308652472005644, 0.7338255988615702], "95": [0.6200345849802371, 0.7411211707957673]}, "precision": 0.768361581920904, "recall": 0.6181818181818182}}}, "2021.test (span detection)": {"micro/f1": 0.7953163189905075, "micro/f1_ci": {}, "micro/recall": 0.8090667283450907, "micro/precision": 0.7820254862508383, "macro/f1": 0.7953163189905075, "macro/f1_ci": {}, "macro/recall": 0.8090667283450907, "macro/precision": 0.7820254862508383}, "2020.test (span detection)": {"micro/f1": 0.7759088442756376, "micro/f1_ci": {}, "micro/recall": 0.742086144265698, "micro/precision": 0.8129619101762365, "macro/f1": 0.7759088442756376, "macro/f1_ci": {}, "macro/recall": 0.742086144265698, "macro/precision": 0.8129619101762365}}

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:6578562cd714d5ac36c62f4c46b4f5854dad90303a3dca5f1830233752a3c42e
-size 1417461873
+oid sha256:370726c3544f72ee7c98940a7ff87c027812a4c30d8ce4a1bfe03454174fc6fb
+size 1417467377

tokenizer_config.json

@@ -1 +1 @@
-{"errors": "replace", "bos_token": "<s>", "eos_token": "</s>", "sep_token": "</s>", "cls_token": "<s>", "unk_token": "<unk>", "pad_token": "<pad>", "mask_token": "<mask>", "add_prefix_space": false, "trim_offsets": true, "name_or_path": "cner_output/model/baseline_2021/bertweet_large_concat/best_model", "special_tokens_map_file": "cner_output/model/baseline_2021/bertweet_large_concat/best_model/special_tokens_map.json", "tokenizer_class": "RobertaTokenizer"}
+{"errors": "replace", "bos_token": "<s>", "eos_token": "</s>", "sep_token": "</s>", "cls_token": "<s>", "unk_token": "<unk>", "pad_token": "<pad>", "mask_token": "<mask>", "add_prefix_space": false, "trim_offsets": true, "name_or_path": "vinai/bertweet-large", "tokenizer_class": "RobertaTokenizer"}

trainer_config.json

@@ -0,0 +1 @@
+{"data_split": "2020_2021.train", "model": "vinai/bertweet-large", "crf": true, "max_length": 128, "epoch": 30, "batch_size": 32, "lr": 1e-05, "random_seed": 0, "gradient_accumulation_steps": 1, "weight_decay": 1e-07, "lr_warmup_step_ratio": 0.15, "max_grad_norm": 1}