model update

README.md

@@ -0,0 +1,144 @@
+---
+datasets:
+- mit_movie_trivia
+metrics:
+- f1
+- precision
+- recall
+model-index:
+- name: tner/deberta-v3-large-mit-movie-trivia
+  results:
+  - task:
+      name: Token Classification
+      type: token-classification
+    dataset:
+      name: mit_movie_trivia
+      type: mit_movie_trivia
+      args: mit_movie_trivia
+    metrics:
+    - name: F1
+      type: f1
+      value: 0.7324478178368122
+    - name: Precision
+      type: precision
+      value: 0.7186865267433988
+    - name: Recall
+      type: recall
+      value: 0.746746394653535
+    - name: F1 (macro)
+      type: f1_macro
+      value: 0.6597589403836301
+    - name: Precision (macro)
+      type: precision_macro
+      value: 0.6493939604029393
+    - name: Recall (macro)
+      type: recall_macro
+      value: 0.6747458149186768
+    - name: F1 (entity span)
+      type: f1_entity_span
+      value: 0.749525289142068
+    - name: Precision (entity span)
+      type: precision_entity_span
+      value: 0.7359322033898306
+    - name: Recall (entity span)
+      type: recall_entity_span
+      value: 0.7636299683432993
+
+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-movie-trivia
+
+This model is a fine-tuned version of [microsoft/deberta-v3-large](https://huggingface.co/microsoft/deberta-v3-large) on the 
+[tner/mit_movie_trivia](https://huggingface.co/datasets/tner/mit_movie_trivia) 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): 0.7324478178368122
+- Precision (micro): 0.7186865267433988
+- Recall (micro): 0.746746394653535
+- F1 (macro): 0.6597589403836301
+- Precision (macro): 0.6493939604029393
+- Recall (macro): 0.6747458149186768
+
+The per-entity breakdown of the F1 score on the test set are below:
+- actor: 0.9590417310664605
+- award: 0.4755244755244755
+- character_name: 0.7391304347826086
+- date: 0.9640179910044978
+- director: 0.909706546275395
+- genre: 0.755114693118413
+- opinion: 0.4910714285714286
+- origin: 0.3922518159806296
+- plot: 0.4929757343550447
+- quote: 0.7391304347826088
+- relationship: 0.5705705705705706
+- soundtrack: 0.42857142857142855 
+
+For F1 scores, the confidence interval is obtained by bootstrap as below:
+- F1 (micro): 
+    - 90%: [0.7213456287685677, 0.742502895519075]
+    - 95%: [0.7198169787204788, 0.7460320515170399] 
+- F1 (macro): 
+    - 90%: [0.7213456287685677, 0.742502895519075]
+    - 95%: [0.7198169787204788, 0.7460320515170399] 
+
+Full evaluation can be found at [metric file of NER](https://huggingface.co/tner/deberta-v3-large-mit-movie-trivia/raw/main/eval/metric.json) 
+and [metric file of entity span](https://huggingface.co/tner/deberta-v3-large-mit-movie-trivia/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-movie-trivia")
+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_movie_trivia']
+ - 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-movie-trivia/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_movie_trivia_deberta_v3_large/best_model",
+  "_name_or_path": "tner_ckpt/mit_movie_trivia_deberta_v3_large/model_rgwuwr/epoch_5",
   "architectures": [
     "DebertaV2ForTokenClassification"
   ],

eval/metric.json

@@ -0,0 +1 @@
+{"micro/f1": 0.7324478178368122, "micro/f1_ci": {"90": [0.7213456287685677, 0.742502895519075], "95": [0.7198169787204788, 0.7460320515170399]}, "micro/recall": 0.746746394653535, "micro/precision": 0.7186865267433988, "macro/f1": 0.6597589403836301, "macro/f1_ci": {"90": [0.626187284258936, 0.6886112877936686], "95": [0.6207351107382689, 0.6929921662444682]}, "macro/recall": 0.6747458149186768, "macro/precision": 0.6493939604029393, "per_entity_metric": {"actor": {"f1": 0.9590417310664605, "f1_ci": {"90": [0.9506119297682675, 0.9674256910027863], "95": [0.9486777930500124, 0.9687562102543722]}, "precision": 0.9444444444444444, "recall": 0.9740973312401884}, "award": {"f1": 0.4755244755244755, "f1_ci": {"90": [0.3902439024390244, 0.5663864306784661], "95": [0.37141705069124425, 0.5828690476190477]}, "precision": 0.44155844155844154, "recall": 0.5151515151515151}, "character_name": {"f1": 0.7391304347826086, "f1_ci": {"90": [0.6994571044586552, 0.7751456876456877], "95": [0.687940448046831, 0.7833757784108667]}, "precision": 0.7015873015873015, "recall": 0.7809187279151943}, "date": {"f1": 0.9640179910044978, "f1_ci": {"90": [0.9531357020547946, 0.9742035620393553], "95": [0.9506469700990182, 0.9770297199170713]}, "precision": 0.9554234769687965, "recall": 0.972768532526475}, "director": {"f1": 0.909706546275395, "f1_ci": {"90": [0.8929881320434827, 0.9263193657984146], "95": [0.88940053548902, 0.9288156974032032]}, "precision": 0.8741865509761388, "recall": 0.9482352941176471}, "genre": {"f1": 0.755114693118413, "f1_ci": {"90": [0.7313228085387712, 0.7795221322448262], "95": [0.7266076737272389, 0.7837107883905914]}, "precision": 0.7390776699029126, "recall": 0.7718631178707225}, "opinion": {"f1": 0.4910714285714286, "f1_ci": {"90": [0.4395577262243929, 0.539553444699203], "95": [0.42980954097572693, 0.5492122875620395]}, "precision": 0.43478260869565216, "recall": 0.5641025641025641}, "origin": {"f1": 0.3922518159806296, "f1_ci": {"90": [0.3405250077663871, 0.44771792935648963], "95": [0.3317985893416927, 0.45755013816329176]}, "precision": 0.3632286995515695, "recall": 0.4263157894736842}, "plot": {"f1": 0.4929757343550447, "f1_ci": {"90": [0.47077832785029755, 0.5149341008990155], "95": [0.4652040411165297, 0.5178380252221502]}, "precision": 0.4964630225080386, "recall": 0.4895370957514268}, "quote": {"f1": 0.7391304347826088, "f1_ci": {"90": [0.6315315315315315, 0.8395308641975308], "95": [0.6105070603337611, 0.8536712398373985]}, "precision": 0.7555555555555555, "recall": 0.723404255319149}, "relationship": {"f1": 0.5705705705705706, "f1_ci": {"90": [0.5073746312684367, 0.6306780464675202], "95": [0.4984200575683226, 0.6439731995821846]}, "precision": 0.5864197530864198, "recall": 0.5555555555555556}, "soundtrack": {"f1": 0.42857142857142855, "f1_ci": {"90": [0.1111111111111111, 0.7142857142857143], "95": [0.0, 0.7777777777777777]}, "precision": 0.5, "recall": 0.375}}}

eval/metric_span.json

@@ -0,0 +1 @@
+{"micro/f1": 0.749525289142068, "micro/f1_ci": {"90": [0.7397126154825829, 0.7597425112279943], "95": [0.7379706607775025, 0.7623178184832154]}, "micro/recall": 0.7636299683432993, "micro/precision": 0.7359322033898306, "macro/f1": 0.749525289142068, "macro/f1_ci": {"90": [0.7397126154825829, 0.7597425112279943], "95": [0.7379706607775025, 0.7623178184832154]}, "macro/recall": 0.7636299683432993, "macro/precision": 0.7359322033898306}

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:8fff54b58199a785de8bcbf0fe86a1f54365a86f91c5acbd6f329d2ca9d4d96d
-size 1736283119
+oid sha256:15087b08ab22212d301c05abd689ed9d798c2cab47d7e3c595f5bd0d409609e7
+size 1736288623

tokenizer_config.json

@@ -4,7 +4,7 @@
   "do_lower_case": false,
   "eos_token": "[SEP]",
   "mask_token": "[MASK]",
-  "name_or_path": "tner_ckpt/mit_movie_trivia_deberta_v3_large/best_model",
+  "name_or_path": "tner_ckpt/mit_movie_trivia_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_movie_trivia"], "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}