Jorge Lopez Grisman
update model card README.md
b566c06
|
raw
history blame
2.35 kB
metadata
license: apache-2.0
tags:
  - generated_from_trainer
datasets:
  - conll2003
metrics:
  - precision
  - recall
  - f1
  - accuracy
model-index:
  - name: albert-base-v2-finetuned-ner
    results:
      - task:
          name: Token Classification
          type: token-classification
        dataset:
          name: conll2003
          type: conll2003
          args: conll2003
        metrics:
          - name: Precision
            type: precision
            value: 0.9252213840603477
          - name: Recall
            type: recall
            value: 0.9329732113328189
          - name: F1
            type: f1
            value: 0.9290811285541773
          - name: Accuracy
            type: accuracy
            value: 0.9848205157332728

albert-base-v2-finetuned-ner

This model is a fine-tuned version of albert-base-v2 on the conll2003 dataset. It achieves the following results on the evaluation set:

  • Loss: 0.0626
  • Precision: 0.9252
  • Recall: 0.9330
  • F1: 0.9291
  • Accuracy: 0.9848

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: 64
  • eval_batch_size: 64
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • num_epochs: 5

Training results

Training Loss Epoch Step Validation Loss Precision Recall F1 Accuracy
No log 1.0 220 0.0863 0.8827 0.8969 0.8898 0.9773
No log 2.0 440 0.0652 0.8951 0.9199 0.9073 0.9809
0.1243 3.0 660 0.0626 0.9191 0.9208 0.9200 0.9827
0.1243 4.0 880 0.0585 0.9227 0.9281 0.9254 0.9843
0.0299 5.0 1100 0.0626 0.9252 0.9330 0.9291 0.9848

Framework versions

  • Transformers 4.16.2
  • Pytorch 1.8.1+cu111
  • Datasets 1.18.3
  • Tokenizers 0.11.0