terzimert's picture
update model card README.md
51698ff
metadata
license: apache-2.0
tags:
  - generated_from_trainer
datasets:
  - caner
metrics:
  - precision
  - recall
  - f1
  - accuracy
model-index:
  - name: bert-finetuned-ner-v4.001
    results:
      - task:
          name: Token Classification
          type: token-classification
        dataset:
          name: caner
          type: caner
          config: default
          split: train[-1%:]
          args: default
        metrics:
          - name: Precision
            type: precision
            value: 0.8814432989690721
          - name: Recall
            type: recall
            value: 0.8208
          - name: F1
            type: f1
            value: 0.8500414250207124
          - name: Accuracy
            type: accuracy
            value: 0.9327371695178849

bert-finetuned-ner-v4.001

This model is a fine-tuned version of bert-base-multilingual-cased on the caner dataset. It achieves the following results on the evaluation set:

  • Loss: 0.4995
  • Precision: 0.8814
  • Recall: 0.8208
  • F1: 0.8500
  • Accuracy: 0.9327

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

  • learning_rate: 2e-05
  • train_batch_size: 8
  • eval_batch_size: 8
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • num_epochs: 3

Training results

Training Loss Epoch Step Validation Loss Precision Recall F1 Accuracy
0.2518 1.0 4842 0.5403 0.8354 0.7712 0.8020 0.9178
0.1364 2.0 9684 0.4746 0.8728 0.8016 0.8357 0.9287
0.0915 3.0 14526 0.4995 0.8814 0.8208 0.8500 0.9327

Framework versions

  • Transformers 4.27.4
  • Pytorch 1.13.1+cu116
  • Datasets 2.11.0
  • Tokenizers 0.13.2