model update

README.md

@@ -0,0 +1,140 @@
+---
+datasets:
+- mit_restaurant
+metrics:
+- f1
+- precision
+- recall
+model-index:
+- name: tner/deberta-v3-large-mit-restaurant
+  results:
+  - task:
+      name: Token Classification
+      type: token-classification
+    dataset:
+      name: mit_restaurant
+      type: mit_restaurant
+      args: mit_restaurant
+    metrics:
+    - name: F1
+      type: f1
+      value: 0.8158890290037831
+    - name: Precision
+      type: precision
+      value: 0.8105230191042906
+    - name: Recall
+      type: recall
+      value: 0.8213265629958744
+    - name: F1 (macro)
+      type: f1_macro
+      value: 0.8072607717138172
+    - name: Precision (macro)
+      type: precision_macro
+      value: 0.7973293573334044
+    - name: Recall (macro)
+      type: recall_macro
+      value: 0.8183493118743246
+    - name: F1 (entity span)
+      type: f1_entity_span
+      value: 0.8557510999371464
+    - name: Precision (entity span)
+      type: precision_entity_span
+      value: 0.8474945533769063
+    - name: Recall (entity span)
+      type: recall_entity_span
+      value: 0.8641701047286575
+
+pipeline_tag: token-classification
+widget:
+- text: "Jacob Collier is a Grammy awarded artist from England."
+  example_title: "NER Example 1"
+---
+# tner/deberta-v3-large-mit-restaurant
+
+This model is a fine-tuned version of [microsoft/deberta-v3-large](https://huggingface.co/microsoft/deberta-v3-large) on the 
+[tner/mit_restaurant](https://huggingface.co/datasets/tner/mit_restaurant) 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): 0.8158890290037831
+- Precision (micro): 0.8105230191042906
+- Recall (micro): 0.8213265629958744
+- F1 (macro): 0.8072607717138172
+- Precision (macro): 0.7973293573334044
+- Recall (macro): 0.8183493118743246
+
+The per-entity breakdown of the F1 score on the test set are below:
+- amenity: 0.7226415094339623
+- cuisine: 0.8288119738072967
+- dish: 0.8283828382838284
+- location: 0.8662969808995686
+- money: 0.84
+- rating: 0.7990430622009569
+- restaurant: 0.8724489795918368
+- time: 0.7004608294930875 
+
+For F1 scores, the confidence interval is obtained by bootstrap as below:
+- F1 (micro): 
+    - 90%: [0.8036180555961564, 0.8281173227233776]
+    - 95%: [0.8011397826491581, 0.8307029010155984] 
+- F1 (macro): 
+    - 90%: [0.8036180555961564, 0.8281173227233776]
+    - 95%: [0.8011397826491581, 0.8307029010155984] 
+
+Full evaluation can be found at [metric file of NER](https://huggingface.co/tner/deberta-v3-large-mit-restaurant/raw/main/eval/metric.json) 
+and [metric file of entity span](https://huggingface.co/tner/deberta-v3-large-mit-restaurant/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/deberta-v3-large-mit-restaurant")
+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/mit_restaurant']
+ - 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: None
+
+The full configuration can be found at [fine-tuning parameter file](https://huggingface.co/tner/deberta-v3-large-mit-restaurant/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/mit_restaurant_deberta_v3_large/best_model",
+  "_name_or_path": "tner_ckpt/mit_restaurant_deberta_v3_large/model_rgwuwr/epoch_5",
   "architectures": [
     "DebertaV2ForTokenClassification"
   ],

eval/metric.json

@@ -0,0 +1 @@
+{"micro/f1": 0.8158890290037831, "micro/f1_ci": {"90": [0.8036180555961564, 0.8281173227233776], "95": [0.8011397826491581, 0.8307029010155984]}, "micro/recall": 0.8213265629958744, "micro/precision": 0.8105230191042906, "macro/f1": 0.8072607717138172, "macro/f1_ci": {"90": [0.7937247965875296, 0.8198696966745797], "95": [0.7907943777789816, 0.8235294293836553]}, "macro/recall": 0.8183493118743246, "macro/precision": 0.7973293573334044, "per_entity_metric": {"amenity": {"f1": 0.7226415094339623, "f1_ci": {"90": [0.6933268608414239, 0.7527957577082878], "95": [0.6885539660746712, 0.7575899667952976]}, "precision": 0.7267552182163188, "recall": 0.7185741088180112}, "cuisine": {"f1": 0.8288119738072967, "f1_ci": {"90": [0.8045715060269436, 0.8507049853075287], "95": [0.800355515041021, 0.8547355413126947]}, "precision": 0.8249534450651769, "recall": 0.8327067669172933}, "dish": {"f1": 0.8283828382838284, "f1_ci": {"90": [0.7977736928104575, 0.8585744093773348], "95": [0.7932086894586895, 0.8630655911732973]}, "precision": 0.789308176100629, "recall": 0.8715277777777778}, "location": {"f1": 0.8662969808995686, "f1_ci": {"90": [0.846563505906978, 0.8860332482724753], "95": [0.84335027383273, 0.8891798434189129]}, "precision": 0.8668310727496917, "recall": 0.8657635467980296}, "money": {"f1": 0.84, "f1_ci": {"90": [0.795573832245103, 0.8812646349862259], "95": [0.7838484630163304, 0.8919160231660233]}, "precision": 0.8212290502793296, "recall": 0.8596491228070176}, "rating": {"f1": 0.7990430622009569, "f1_ci": {"90": [0.7589409831260728, 0.83568415322855], "95": [0.7492980278849697, 0.8430649055847957]}, "precision": 0.7695852534562212, "recall": 0.8308457711442786}, "restaurant": {"f1": 0.8724489795918368, "f1_ci": {"90": [0.8464719582385224, 0.897852277205427], "95": [0.8424846340100577, 0.9029754204398447]}, "precision": 0.8952879581151832, "recall": 0.8507462686567164}, "time": {"f1": 0.7004608294930875, "f1_ci": {"90": [0.6505488597424081, 0.7458471061796478], "95": [0.6403466036873743, 0.7572871589049083]}, "precision": 0.6846846846846847, "recall": 0.7169811320754716}}}

eval/metric_span.json

@@ -0,0 +1 @@
+{"micro/f1": 0.8557510999371464, "micro/f1_ci": {"90": [0.8457549886322284, 0.866532074562069], "95": [0.8439094859106241, 0.8689014756604962]}, "micro/recall": 0.8641701047286575, "micro/precision": 0.8474945533769063, "macro/f1": 0.8557510999371464, "macro/f1_ci": {"90": [0.8457549886322284, 0.866532074562069], "95": [0.8439094859106241, 0.8689014756604962]}, "macro/recall": 0.8641701047286575, "macro/precision": 0.8474945533769063}

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:66c995d18c8a759eacf6d17a8ca16d349899fa047816497bfe14c99950680dcd
-size 1736250351
+oid sha256:b23b1c8747da48934080fc97c1c41b16f6ad5846bfcc347b5b67f519bce0f1b8
+size 1736255855

tokenizer_config.json

@@ -4,7 +4,7 @@
   "do_lower_case": false,
   "eos_token": "[SEP]",
   "mask_token": "[MASK]",
-  "name_or_path": "tner_ckpt/mit_restaurant_deberta_v3_large/best_model",
+  "name_or_path": "tner_ckpt/mit_restaurant_deberta_v3_large/model_rgwuwr/epoch_5",
   "pad_token": "[PAD]",
   "sep_token": "[SEP]",
   "sp_model_kwargs": {},

trainer_config.json

@@ -0,0 +1 @@
+{"dataset": ["tner/mit_restaurant"], "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": null}