Endika99/NLP-TokenClass-NER

1---2license: apache-2.03tags:4- generated_from_trainer5datasets:6- wikiann7metrics:8- precision9- recall10- f111- accuracy12model-index:13- name: NLP-TokenClass-NER14  results:15  - task:16      name: Token Classification17      type: token-classification18    dataset:19      name: wikiann20      type: wikiann21      config: en22      split: validation23      args: en24    metrics:25    - name: Precision26      type: precision27      value: 0.811963482763319228    - name: Recall29      type: recall30      value: 0.842499646543192431    - name: F132      type: f133      value: 0.826949764085484434    - name: Accuracy35      type: accuracy36      value: 0.9254762384830537---38 39<!-- This model card has been generated automatically according to the information the Trainer had access to. You40should probably proofread and complete it, then remove this comment. -->41 42# NLP-TokenClass-NER43 44This model is a fine-tuned version of [bert-base-cased](https://huggingface.co/bert-base-cased) on the wikiann dataset.45It achieves the following results on the evaluation set:46- Loss: 0.374147- Precision: 0.812048- Recall: 0.842549- F1: 0.826950- Accuracy: 0.925551 52## Model description53 54More information needed55 56## Intended uses & limitations57 58More information needed59 60## Training and evaluation data61 62More information needed63 64## Training procedure65 66### Training hyperparameters67 68The following hyperparameters were used during training:69- learning_rate: 2e-0570- train_batch_size: 871- eval_batch_size: 872- seed: 4273- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-0874- lr_scheduler_type: linear75- num_epochs: 176 77### Training results78 79| Training Loss | Epoch | Step | Validation Loss | Precision | Recall | F1     | Accuracy |80|:-------------:|:-----:|:----:|:---------------:|:---------:|:------:|:------:|:--------:|81| 0.0902        | 1.0   | 2500 | 0.3741          | 0.8120    | 0.8425 | 0.8269 | 0.9255   |82 83 84### Framework versions85 86- Transformers 4.26.187- Pytorch 1.13.1+cu11688- Datasets 2.10.189- Tokenizers 0.13.290