model update

README.md

@@ -0,0 +1,126 @@
+---
+datasets:
+- wnut2017
+metrics:
+- f1
+- precision
+- recall
+model-index:
+- name: tner/deberta-large-wnut2017
+  results:
+  - task:
+      name: Token Classification
+      type: token-classification
+    dataset:
+      name: wnut2017
+      type: wnut2017
+      args: wnut2017
+    metrics:
+    - name: F1
+      type: f1
+      value: 0.5105386416861827
+    - name: Precision
+      type: precision
+      value: 0.6931637519872814
+    - name: Recall
+      type: recall
+      value: 0.4040778498609824
+    - name: F1 (macro)
+      type: f1_macro
+      value: 0.4263428845085451
+    - name: Precision (macro)
+      type: precision_macro
+      value: 0.6003185137596864
+    - name: Recall (macro)
+      type: recall_macro
+      value: 0.35195768262641947
+    - name: F1 (entity span)
+      type: f1_entity_span
+      value: 0.5936768149882904
+    - name: Precision (entity span)
+      type: precision_entity_span
+      value: 0.8060413354531002
+    - name: Recall (entity span)
+      type: recall_entity_span
+      value: 0.46987951807228917
+
+pipeline_tag: token-classification
+widget:
+- text: "Jacob Collier is a Grammy awarded artist from England."
+  example_title: "NER Example 1"
+---
+# tner/deberta-large-wnut2017
+
+This model is a fine-tuned version of [microsoft/deberta-large](https://huggingface.co/microsoft/deberta-large) on the 
+[tner/wnut2017](https://huggingface.co/datasets/tner/wnut2017) 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): 0.5105386416861827
+- Precision (micro): 0.6931637519872814
+- Recall (micro): 0.4040778498609824
+- F1 (macro): 0.4263428845085451
+- Precision (macro): 0.6003185137596864
+- Recall (macro): 0.35195768262641947
+
+The per-entity breakdown of the F1 score on the test set are below:
+- corporation: 0.3503649635036496
+- group: 0.3148148148148148
+- location: 0.6029411764705882
+- person: 0.6628895184135977
+- product: 0.1951219512195122
+- work_of_art: 0.431924882629108 
+
+For F1 scores, the confidence interval is obtained by bootstrap as below:
+- F1 (micro): 
+    - 90%: [0.47970650356554456, 0.5385161869734422]
+    - 95%: [0.47475901512925966, 0.5430870496346687] 
+- F1 (macro): 
+    - 90%: [0.47970650356554456, 0.5385161869734422]
+    - 95%: [0.47475901512925966, 0.5430870496346687] 
+
+Full evaluation can be found at [metric file of NER](https://huggingface.co/tner/deberta-large-wnut2017/raw/main/eval/metric.json) 
+and [metric file of entity span](https://huggingface.co/tner/deberta-large-wnut2017/raw/main/eval/metric_span.json).
+
+
+### Training hyperparameters
+
+The following hyperparameters were used during training:
+ - dataset: ['tner/wnut2017']
+ - dataset_split: train
+ - dataset_name: None
+ - local_dataset: None
+ - model: microsoft/deberta-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-large-wnut2017/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/wnut2017_deberta_large/best_model",
+  "_name_or_path": "tner_ckpt/wnut2017_deberta_large_v1/model_ayzfwn/epoch_5",
   "architectures": [
     "DebertaForTokenClassification"
   ],

eval/metric.json

@@ -0,0 +1 @@
+{"micro/f1": 0.5105386416861827, "micro/f1_ci": {"90": [0.47970650356554456, 0.5385161869734422], "95": [0.47475901512925966, 0.5430870496346687]}, "micro/recall": 0.4040778498609824, "micro/precision": 0.6931637519872814, "macro/f1": 0.4263428845085451, "macro/f1_ci": {"90": [0.39426591884558804, 0.4545712655238229], "95": [0.38807589943775705, 0.4599705771824737]}, "macro/recall": 0.35195768262641947, "macro/precision": 0.6003185137596864, "per_entity_metric": {"corporation": {"f1": 0.3503649635036496, "f1_ci": {"90": [0.25, 0.44004724409448814], "95": [0.23425332991803258, 0.4615702479338844]}, "precision": 0.3380281690140845, "recall": 0.36363636363636365}, "group": {"f1": 0.3148148148148148, "f1_ci": {"90": [0.2272727272727273, 0.39614832223527874], "95": [0.21596110755441741, 0.4055479550691245]}, "precision": 0.6666666666666666, "recall": 0.20606060606060606}, "location": {"f1": 0.6029411764705882, "f1_ci": {"90": [0.535979661016949, 0.6666666666666665], "95": [0.5179212561518631, 0.6752398081534774]}, "precision": 0.6721311475409836, "recall": 0.5466666666666666}, "person": {"f1": 0.6628895184135977, "f1_ci": {"90": [0.6241108987808454, 0.6986254012381062], "95": [0.6153846153846154, 0.7045139877293182]}, "precision": 0.8447653429602888, "recall": 0.5454545454545454}, "product": {"f1": 0.1951219512195122, "f1_ci": {"90": [0.12413793103448277, 0.26666666666666666], "95": [0.11031844733371016, 0.28111927144535853]}, "precision": 0.43243243243243246, "recall": 0.12598425196850394}, "work_of_art": {"f1": 0.431924882629108, "f1_ci": {"90": [0.3571260898725687, 0.5065021645021645], "95": [0.34593366093366096, 0.5167490995107791]}, "precision": 0.647887323943662, "recall": 0.323943661971831}}}

eval/metric_span.json

@@ -0,0 +1 @@
+{"micro/f1": 0.5936768149882904, "micro/f1_ci": {"90": [0.5661684827134196, 0.6210986510760949], "95": [0.5605935659555948, 0.6248693729903536]}, "micro/recall": 0.46987951807228917, "micro/precision": 0.8060413354531002, "macro/f1": 0.5936768149882904, "macro/f1_ci": {"90": [0.5661684827134196, 0.6210986510760949], "95": [0.5605935659555948, 0.6248693729903536]}, "macro/recall": 0.46987951807228917, "macro/precision": 0.8060413354531002}

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:34c3badba8b88657cab3d60a67e0bc7e0253de374586436f11839e5bfcff6d56
-size 1620836275
+oid sha256:9ef61217f4af95ab9b1640b0af9790ebc770bb933e5ebab187f2120e711e33d1
+size 1620841715

tokenizer_config.json

@@ -36,7 +36,7 @@
     "single_word": false
   },
   "model_max_length": 512,
-  "name_or_path": "tner_ckpt/wnut2017_deberta_large/best_model",
+  "name_or_path": "tner_ckpt/wnut2017_deberta_large_v1/model_ayzfwn/epoch_5",
   "pad_token": {
     "__type": "AddedToken",
     "content": "[PAD]",

trainer_config.json

@@ -0,0 +1 @@
+{"dataset": ["tner/wnut2017"], "dataset_split": "train", "dataset_name": null, "local_dataset": null, "model": "microsoft/deberta-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}