model update

README.md

@@ -0,0 +1,142 @@
+---
+datasets:
+- tweebank_ner
+metrics:
+- f1
+- precision
+- recall
+model-index:
+- name: tner/deberta-v3-large-tweebank-ner
+  results:
+  - task:
+      name: Token Classification
+      type: token-classification
+    dataset:
+      name: tweebank_ner
+      type: tweebank_ner
+      args: tweebank_ner
+    metrics:
+    - name: F1
+      type: f1
+      value: 0.7253474520185308
+    - name: Precision
+      type: precision
+      value: 0.7201051248357424
+    - name: Recall
+      type: recall
+      value: 0.7306666666666667
+    - name: F1 (macro)
+      type: f1_macro
+      value: 0.701874697798745
+    - name: Precision (macro)
+      type: precision_macro
+      value: 0.7043005470796733
+    - name: Recall (macro)
+      type: recall_macro
+      value: 0.706915721861374
+    - name: F1 (entity span)
+      type: f1_entity_span
+      value: 0.8178343949044585
+    - name: Precision (entity span)
+      type: precision_entity_span
+      value: 0.7829268292682927
+    - name: Recall (entity span)
+      type: recall_entity_span
+      value: 0.856
+
+pipeline_tag: token-classification
+widget:
+- text: "Jacob Collier is a Grammy awarded artist from England."
+  example_title: "NER Example 1"
+---
+# tner/deberta-v3-large-tweebank-ner
+
+This model is a fine-tuned version of [microsoft/deberta-v3-large](https://huggingface.co/microsoft/deberta-v3-large) on the 
+[tner/tweebank_ner](https://huggingface.co/datasets/tner/tweebank_ner) 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): 0.7253474520185308
+- Precision (micro): 0.7201051248357424
+- Recall (micro): 0.7306666666666667
+- F1 (macro): 0.701874697798745
+- Precision (macro): 0.7043005470796733
+- Recall (macro): 0.706915721861374
+
+The per-entity breakdown of the F1 score on the test set are below:
+- location: 0.7289719626168224
+- organization: 0.7040816326530612
+- other: 0.5182926829268293
+- person: 0.856152512998267 
+
+For F1 scores, the confidence interval is obtained by bootstrap as below:
+- F1 (micro): 
+    - 90%: [0.6978100031831928, 0.7529703029130037]
+    - 95%: [0.691700704571692, 0.7582901338971108] 
+- F1 (macro): 
+    - 90%: [0.6978100031831928, 0.7529703029130037]
+    - 95%: [0.691700704571692, 0.7582901338971108] 
+
+Full evaluation can be found at [metric file of NER](https://huggingface.co/tner/deberta-v3-large-tweebank-ner/raw/main/eval/metric.json) 
+and [metric file of entity span](https://huggingface.co/tner/deberta-v3-large-tweebank-ner/raw/main/eval/metric_span.json).
+
+### Usage
+This model can be used through the transformers library by 
+```python
+from transformers import AutoTokenizer, AutoModelForTokenClassification
+tokenizer = AutoTokenizer.from_pretrained("tner/deberta-v3-large-tweebank-ner")
+model = AutoModelForTokenClassification.from_pretrained("tner/deberta-v3-large-tweebank-ner")
+```
+but, since transformers do not support CRF layer, it is recommended to use the model via `T-NER` library. 
+Install the library via pip   
+```shell
+pip install tner
+```
+and activate model as below.
+```python
+from tner import TransformersNER
+model = TransformersNER("tner/deberta-v3-large-tweebank-ner")
+model.predict("Jacob Collier is a Grammy awarded English artist from London".split(" "))
+```
+
+### Training hyperparameters
+
+The following hyperparameters were used during training:
+ - dataset: ['tner/tweebank_ner']
+ - dataset_split: train
+ - dataset_name: None
+ - local_dataset: None
+ - model: microsoft/deberta-v3-large
+ - crf: True
+ - max_length: 128
+ - epoch: 15
+ - batch_size: 16
+ - lr: 1e-05
+ - random_seed: 42
+ - gradient_accumulation_steps: 4
+ - weight_decay: 1e-07
+ - 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/deberta-v3-large-tweebank-ner/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/tweebank_ner_deberta_v3_large/best_model",
+  "_name_or_path": "tner_ckpt/tweebank_ner_deberta_v3_large/model_ayzfwn/epoch_5",
   "architectures": [
     "DebertaV2ForTokenClassification"
   ],

eval/metric.json

@@ -0,0 +1 @@
+{"micro/f1": 0.7253474520185308, "micro/f1_ci": {"90": [0.6978100031831928, 0.7529703029130037], "95": [0.691700704571692, 0.7582901338971108]}, "micro/recall": 0.7306666666666667, "micro/precision": 0.7201051248357424, "macro/f1": 0.701874697798745, "macro/f1_ci": {"90": [0.6718178675509049, 0.7319530846059807], "95": [0.6668177051487543, 0.7376052570235649]}, "macro/recall": 0.706915721861374, "macro/precision": 0.7043005470796733, "per_entity_metric": {"location": {"f1": 0.7289719626168224, "f1_ci": {"90": [0.6562417328042328, 0.7966673014663873], "95": [0.641968706674589, 0.8097933196199433]}, "precision": 0.7572815533980582, "recall": 0.7027027027027027}, "organization": {"f1": 0.7040816326530612, "f1_ci": {"90": [0.654310820624546, 0.7532017312488011], "95": [0.6434627329192546, 0.7627998776009792]}, "precision": 0.6571428571428571, "recall": 0.7582417582417582}, "other": {"f1": 0.5182926829268293, "f1_ci": {"90": [0.4594546843663452, 0.5747992256531156], "95": [0.44514336917562725, 0.5848251201251539]}, "precision": 0.5902777777777778, "recall": 0.46195652173913043}, "person": {"f1": 0.856152512998267, "f1_ci": {"90": [0.8253399015339731, 0.8826298839871183], "95": [0.8206082403601488, 0.8872012145386544]}, "precision": 0.8125, "recall": 0.9047619047619048}}}

eval/metric_span.json

@@ -0,0 +1 @@
+{"micro/f1": 0.8178343949044585, "micro/f1_ci": {"90": [0.7946512020674701, 0.8391476596836752], "95": [0.790463656866032, 0.8430815891456205]}, "micro/recall": 0.856, "micro/precision": 0.7829268292682927, "macro/f1": 0.8178343949044585, "macro/f1_ci": {"90": [0.7946512020674701, 0.8391476596836752], "95": [0.790463656866032, 0.8430815891456205]}, "macro/recall": 0.856, "macro/precision": 0.7829268292682927}

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:8e6734e5a6341e0f6ee6e22d35485a57b83d4461a5bb4bb7f78bd5d8f85bedc9
-size 1736217519
+oid sha256:35f73ab3f2b995a6b0a3d5df48a169408d614a6ed5502fd00ac1ae313ce2faa9
+size 1736223023

tokenizer_config.json

@@ -4,7 +4,7 @@
   "do_lower_case": false,
   "eos_token": "[SEP]",
   "mask_token": "[MASK]",
-  "name_or_path": "tner_ckpt/tweebank_ner_deberta_v3_large/best_model",
+  "name_or_path": "tner_ckpt/tweebank_ner_deberta_v3_large/model_ayzfwn/epoch_5",
   "pad_token": "[PAD]",
   "sep_token": "[SEP]",
   "sp_model_kwargs": {},

trainer_config.json

@@ -0,0 +1 @@
+{"dataset": ["tner/tweebank_ner"], "dataset_split": "train", "dataset_name": null, "local_dataset": null, "model": "microsoft/deberta-v3-large", "crf": true, "max_length": 128, "epoch": 15, "batch_size": 16, "lr": 1e-05, "random_seed": 42, "gradient_accumulation_steps": 4, "weight_decay": 1e-07, "lr_warmup_step_ratio": 0.1, "max_grad_norm": 10.0}