model update

README.md

@@ -0,0 +1,150 @@
+---
+datasets:
+- ontonotes5
+metrics:
+- f1
+- precision
+- recall
+model-index:
+- name: tner/roberta-large-ontonotes5
+  results:
+  - task:
+      name: Token Classification
+      type: token-classification
+    dataset:
+      name: ontonotes5
+      type: ontonotes5
+      args: ontonotes5
+    metrics:
+    - name: F1
+      type: f1
+      value: 0.908632361399938
+    - name: Precision
+      type: precision
+      value: 0.905148095909732
+    - name: Recall
+      type: recall
+      value: 0.9121435551212579
+    - name: F1 (macro)
+      type: f1_macro
+      value: 0.8265477704565624
+    - name: Precision (macro)
+      type: precision_macro
+      value: 0.8170668848546687
+    - name: Recall (macro)
+      type: recall_macro
+      value: 0.8387672780349001
+    - name: F1 (entity span)
+      type: f1_entity_span
+      value: 0.9284544931640193
+    - name: Precision (entity span)
+      type: precision_entity_span
+      value: 0.9248942172073342
+    - name: Recall (entity span)
+      type: recall_entity_span
+      value: 0.9320422848005685
+
+pipeline_tag: token-classification
+widget:
+- text: "Jacob Collier is a Grammy awarded artist from England."
+  example_title: "NER Example 1"
+---
+# tner/roberta-large-ontonotes5
+
+This model is a fine-tuned version of [roberta-large](https://huggingface.co/roberta-large) on the 
+[tner/ontonotes5](https://huggingface.co/datasets/tner/ontonotes5) dataset.
+Model fine-tuning is done via [T-NER](https://github.com/asahi417/tner)'s hyper-parameter search (see the repository
+for more detail). It achieves the following results on the test set:
+- F1 (micro): 0.908632361399938
+- Precision (micro): 0.905148095909732
+- Recall (micro): 0.9121435551212579
+- F1 (macro): 0.8265477704565624
+- Precision (macro): 0.8170668848546687
+- Recall (macro): 0.8387672780349001
+
+The per-entity breakdown of the F1 score on the test set are below:
+- cardinal_number: 0.8605277329025309
+- date: 0.872996300863132
+- event: 0.7424242424242424
+- facility: 0.7732342007434945
+- geopolitical_area: 0.9687148323205043
+- group: 0.9470588235294117
+- language: 0.7499999999999999
+- law: 0.6666666666666666
+- location: 0.7593582887700535
+- money: 0.901098901098901
+- ordinal_number: 0.85785536159601
+- organization: 0.9227360841872057
+- percent: 0.9171428571428571
+- person: 0.9556004036326943
+- product: 0.7857142857142858
+- quantity: 0.7945205479452055
+- time: 0.6870588235294116
+- work_of_art: 0.7151515151515151 
+
+For F1 scores, the confidence interval is obtained by bootstrap as below:
+- F1 (micro): 
+    - 90%: [0.9039454247544766, 0.9128956119702822]
+    - 95%: [0.9030263216115454, 0.9138350859566045] 
+- F1 (macro): 
+    - 90%: [0.9039454247544766, 0.9128956119702822]
+    - 95%: [0.9030263216115454, 0.9138350859566045] 
+
+Full evaluation can be found at [metric file of NER](https://huggingface.co/tner/roberta-large-ontonotes5/raw/main/eval/metric.json) 
+and [metric file of entity span](https://huggingface.co/tner/roberta-large-ontonotes5/raw/main/eval/metric_span.json).
+
+### Usage
+This model can be used through the [tner library](https://github.com/asahi417/tner). Install the library via pip   
+```shell
+pip install tner
+```
+and activate model as below.
+```python
+from tner import TransformersNER
+model = TransformersNER("tner/roberta-large-ontonotes5")
+model.predict(["Jacob Collier is a Grammy awarded English artist from London"])
+```
+It can be used via transformers library but it is not recommended as CRF layer is not supported at the moment.
+
+### Training hyperparameters
+
+The following hyperparameters were used during training:
+ - dataset: ['tner/ontonotes5']
+ - dataset_split: train
+ - dataset_name: None
+ - local_dataset: None
+ - model: roberta-large
+ - crf: True
+ - max_length: 128
+ - epoch: 15
+ - batch_size: 64
+ - lr: 1e-05
+ - random_seed: 42
+ - gradient_accumulation_steps: 1
+ - weight_decay: None
+ - lr_warmup_step_ratio: 0.1
+ - max_grad_norm: 10.0
+
+The full configuration can be found at [fine-tuning parameter file](https://huggingface.co/tner/roberta-large-ontonotes5/raw/main/trainer_config.json).
+
+### Reference
+If you use any resource from T-NER, please consider to cite our [paper](https://aclanthology.org/2021.eacl-demos.7/).
+
+```
+
+@inproceedings{ushio-camacho-collados-2021-ner,
+    title = "{T}-{NER}: An All-Round Python Library for Transformer-based Named Entity Recognition",
+    author = "Ushio, Asahi  and
+      Camacho-Collados, Jose",
+    booktitle = "Proceedings of the 16th Conference of the European Chapter of the Association for Computational Linguistics: System Demonstrations",
+    month = apr,
+    year = "2021",
+    address = "Online",
+    publisher = "Association for Computational Linguistics",
+    url = "https://aclanthology.org/2021.eacl-demos.7",
+    doi = "10.18653/v1/2021.eacl-demos.7",
+    pages = "53--62",
+    abstract = "Language model (LM) pretraining has led to consistent improvements in many NLP downstream tasks, including named entity recognition (NER). In this paper, we present T-NER (Transformer-based Named Entity Recognition), a Python library for NER LM finetuning. In addition to its practical utility, T-NER facilitates the study and investigation of the cross-domain and cross-lingual generalization ability of LMs finetuned on NER. Our library also provides a web app where users can get model predictions interactively for arbitrary text, which facilitates qualitative model evaluation for non-expert programmers. We show the potential of the library by compiling nine public NER datasets into a unified format and evaluating the cross-domain and cross- lingual performance across the datasets. The results from our initial experiments show that in-domain performance is generally competitive across datasets. However, cross-domain generalization is challenging even with a large pretrained LM, which has nevertheless capacity to learn domain-specific features if fine- tuned on a combined dataset. To facilitate future research, we also release all our LM checkpoints via the Hugging Face model hub.",
+}
+
+```

config.json

@@ -1,5 +1,5 @@
 {
-  "_name_or_path": "tner_ckpt/ontonotes5_roberta_large/best_model",
+  "_name_or_path": "tner_ckpt/ontonotes5_roberta_large/model_rcsnba/epoch_5",
   "architectures": [
     "RobertaForTokenClassification"
   ],

eval/metric.json

@@ -0,0 +1 @@
+{"micro/f1": 0.908632361399938, "micro/f1_ci": {"90": [0.9039454247544766, 0.9128956119702822], "95": [0.9030263216115454, 0.9138350859566045]}, "micro/recall": 0.9121435551212579, "micro/precision": 0.905148095909732, "macro/f1": 0.8265477704565624, "macro/f1_ci": {"90": [0.8121417948576588, 0.8385364082840391], "95": [0.8096775897485154, 0.8411472970680649]}, "macro/recall": 0.8387672780349001, "macro/precision": 0.8170668848546687, "per_entity_metric": {"cardinal_number": {"f1": 0.8605277329025309, "f1_ci": {"90": [0.843480407184333, 0.8774608056933718], "95": [0.8404683840749414, 0.8804760960168132]}, "precision": 0.8665943600867679, "recall": 0.8545454545454545}, "date": {"f1": 0.872996300863132, "f1_ci": {"90": [0.8597628567219345, 0.8854157808847528], "95": [0.8568743233248947, 0.887132819190551]}, "precision": 0.8623629719853837, "recall": 0.8838951310861424}, "event": {"f1": 0.7424242424242424, "f1_ci": {"90": [0.6564641221374047, 0.8194444444444444], "95": [0.6382887013939839, 0.8345352064804848]}, "precision": 0.7101449275362319, "recall": 0.7777777777777778}, "facility": {"f1": 0.7732342007434945, "f1_ci": {"90": [0.7127132266874755, 0.8299636133523177], "95": [0.6999887892376682, 0.8435233798195244]}, "precision": 0.7761194029850746, "recall": 0.7703703703703704}, "geopolitical_area": {"f1": 0.9687148323205043, "f1_ci": {"90": [0.963340376224374, 0.973711244759882], "95": [0.9626303980811273, 0.9744505463479681]}, "precision": 0.9768497503404449, "recall": 0.9607142857142857}, "group": {"f1": 0.9470588235294117, "f1_ci": {"90": [0.9351070280371709, 0.9580140382076361], "95": [0.932251234137772, 0.9600983382751339]}, "precision": 0.9371362048894063, "recall": 0.9571938168846611}, "language": {"f1": 0.7499999999999999, "f1_ci": {"90": [0.608695652173913, 0.851103230890465], "95": [0.5853150406504064, 0.8666981132075472]}, "precision": 0.8333333333333334, "recall": 0.6818181818181818}, "law": {"f1": 0.6666666666666666, "f1_ci": {"90": [0.5510204081632654, 0.7715238095238094], "95": [0.5318354120811933, 0.7912915851272019]}, "precision": 0.6170212765957447, "recall": 0.725}, "location": {"f1": 0.7593582887700535, "f1_ci": {"90": [0.7075784227499267, 0.8066365067926441], "95": [0.6966818766818768, 0.8132445889832065]}, "precision": 0.7282051282051282, "recall": 0.7932960893854749}, "money": {"f1": 0.901098901098901, "f1_ci": {"90": [0.8723662818004926, 0.9285766271574131], "95": [0.8651037567084079, 0.9333382570162481]}, "precision": 0.8885448916408669, "recall": 0.9140127388535032}, "ordinal_number": {"f1": 0.85785536159601, "f1_ci": {"90": [0.8256974045428536, 0.8873578751164959], "95": [0.8192771084337349, 0.8941270912034538]}, "precision": 0.8349514563106796, "recall": 0.882051282051282}, "organization": {"f1": 0.9227360841872057, "f1_ci": {"90": [0.9125116290604286, 0.9314409091710344], "95": [0.9105165804108345, 0.9326758594060797]}, "precision": 0.9174008810572687, "recall": 0.928133704735376}, "percent": {"f1": 0.9171428571428571, "f1_ci": {"90": [0.888888888888889, 0.9445316700250458], "95": [0.8843537414965987, 0.9478885648316929]}, "precision": 0.9145299145299145, "recall": 0.9197707736389685}, "person": {"f1": 0.9556004036326943, "f1_ci": {"90": [0.949191731901367, 0.9624522822302698], "95": [0.947759511512266, 0.9635008157600174]}, "precision": 0.958502024291498, "recall": 0.9527162977867203}, "product": {"f1": 0.7857142857142858, "f1_ci": {"90": [0.7074746971959516, 0.8571428571428572], "95": [0.6928086318326213, 0.8679526055705302]}, "precision": 0.717391304347826, "recall": 0.868421052631579}, "quantity": {"f1": 0.7945205479452055, "f1_ci": {"90": [0.7301507936507936, 0.8559019931014562], "95": [0.7156685137823244, 0.8688611053468909]}, "precision": 0.7631578947368421, "recall": 0.8285714285714286}, "time": {"f1": 0.6870588235294116, "f1_ci": {"90": [0.6359639016897082, 0.732240357927006], "95": [0.6260340701377503, 0.7433238443393313]}, "precision": 0.6854460093896714, "recall": 0.6886792452830188}, "work_of_art": {"f1": 0.7151515151515151, "f1_ci": {"90": [0.6579761700668609, 0.7656589673913042], "95": [0.6454941116231439, 0.7744146669002161]}, "precision": 0.7195121951219512, "recall": 0.7108433734939759}}}

eval/metric_span.json

@@ -0,0 +1 @@
+{"micro/f1": 0.9284544931640193, "micro/f1_ci": {"90": [0.9241692474050558, 0.9323166980487259], "95": [0.9234176178451158, 0.933219334035083]}, "micro/recall": 0.9320422848005685, "micro/precision": 0.9248942172073342, "macro/f1": 0.9284544931640193, "macro/f1_ci": {"90": [0.9241692474050558, 0.9323166980487259], "95": [0.9234176178451158, 0.933219334035083]}, "macro/recall": 0.9320422848005685, "macro/precision": 0.9248942172073342}

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:cc0b233c63b5c85dbcc016f8db4c7606c4dec4de319051f7a21455cde0059c78
-size 1417523377
+oid sha256:ede8a820e2f09bc0ace94429dd2196a764fa7913196ff27650aa1194c7086131
+size 1417528817

tokenizer_config.json

@@ -6,7 +6,7 @@
   "errors": "replace",
   "mask_token": "<mask>",
   "model_max_length": 512,
-  "name_or_path": "tner_ckpt/ontonotes5_roberta_large/best_model",
+  "name_or_path": "tner_ckpt/ontonotes5_roberta_large/model_rcsnba/epoch_5",
   "pad_token": "<pad>",
   "sep_token": "</s>",
   "special_tokens_map_file": null,

trainer_config.json

@@ -0,0 +1 @@
+{"dataset": ["tner/ontonotes5"], "dataset_split": "train", "dataset_name": null, "local_dataset": null, "model": "roberta-large", "crf": true, "max_length": 128, "epoch": 15, "batch_size": 64, "lr": 1e-05, "random_seed": 42, "gradient_accumulation_steps": 1, "weight_decay": null, "lr_warmup_step_ratio": 0.1, "max_grad_norm": 10.0}