mHossain's picture
Training complete
8791cf1
metadata
base_model: mHossain/bengali_pos_v1_400000
tags:
  - generated_from_trainer
datasets:
  - pos_tag_100k
metrics:
  - precision
  - recall
  - f1
  - accuracy
model-index:
  - name: bengali_pos_v1_500000
    results:
      - task:
          name: Token Classification
          type: token-classification
        dataset:
          name: pos_tag_100k
          type: pos_tag_100k
          config: conll2003
          split: validation
          args: conll2003
        metrics:
          - name: Precision
            type: precision
            value: 0.7517491791092168
          - name: Recall
            type: recall
            value: 0.7532968740056972
          - name: F1
            type: f1
            value: 0.752522230781312
          - name: Accuracy
            type: accuracy
            value: 0.8148741068930406

bengali_pos_v1_500000

This model is a fine-tuned version of mHossain/bengali_pos_v1_400000 on the pos_tag_100k dataset. It achieves the following results on the evaluation set:

  • Loss: 0.6185
  • Precision: 0.7517
  • Recall: 0.7533
  • F1: 0.7525
  • Accuracy: 0.8149

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
0.6508 1.0 67500 0.6325 0.7386 0.7396 0.7391 0.8041
0.5713 2.0 135000 0.6093 0.7487 0.7500 0.7493 0.8123
0.4789 3.0 202500 0.6185 0.7517 0.7533 0.7525 0.8149

Framework versions

  • Transformers 4.35.2
  • Pytorch 2.1.0+cu121
  • Datasets 2.16.1
  • Tokenizers 0.15.0