swtb's picture
End of training
6cec8f0 verified
metadata
license: mit
base_model: xlm-roberta-base
tags:
  - generated_from_trainer
datasets:
  - conll2003
metrics:
  - precision
  - recall
  - f1
  - accuracy
model-index:
  - name: XLM-RoBERTa-Base-Conll2003-English-NER-Finetune-BinaryClass-WeightedLoss
    results:
      - task:
          name: Token Classification
          type: token-classification
        dataset:
          name: conll2003
          type: conll2003
          config: conll2003
          split: test
          args: conll2003
        metrics:
          - name: Precision
            type: precision
            value: 0.952513966480447
          - name: Recall
            type: recall
            value: 0.9660056657223796
          - name: F1
            type: f1
            value: 0.9592123769338959
          - name: Accuracy
            type: accuracy
            value: 0.9906536018089803

XLM-RoBERTa-Base-Conll2003-English-NER-Finetune-BinaryClass-WeightedLoss

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

  • Loss: 0.1110
  • Precision: 0.9525
  • Recall: 0.9660
  • F1: 0.9592
  • Accuracy: 0.9907

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: 5e-06
  • train_batch_size: 4
  • eval_batch_size: 4
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • lr_scheduler_warmup_ratio: 0.05
  • num_epochs: 10

Training results

Training Loss Epoch Step Validation Loss Precision Recall F1 Accuracy
0.3351 0.3333 1441 0.1107 0.9131 0.9377 0.9252 0.9814
0.0357 0.6667 2882 0.0501 0.9500 0.9446 0.9473 0.9882
0.0239 1.0 4323 0.0662 0.9481 0.9479 0.9480 0.9883
0.0178 1.3333 5764 0.0667 0.9489 0.9603 0.9546 0.9899
0.0188 1.6667 7205 0.0712 0.9488 0.9575 0.9531 0.9895
0.018 2.0 8646 0.0605 0.9524 0.9559 0.9541 0.9902
0.0119 2.3333 10087 0.0840 0.9487 0.9662 0.9574 0.9901
0.0124 2.6667 11528 0.0758 0.9486 0.9641 0.9563 0.9901
0.0112 3.0 12969 0.0664 0.9559 0.9628 0.9593 0.9910
0.0082 3.3333 14410 0.0939 0.9483 0.9603 0.9543 0.9899
0.0083 3.6667 15851 0.0681 0.9555 0.9591 0.9573 0.9907
0.0077 4.0 17292 0.0686 0.9555 0.9572 0.9563 0.9902
0.0055 4.3333 18733 0.0852 0.9498 0.9642 0.9569 0.9905
0.005 4.6667 20174 0.0795 0.9530 0.9653 0.9591 0.9907
0.0049 5.0 21615 0.0871 0.9526 0.9614 0.9570 0.9900
0.0042 5.3333 23056 0.1054 0.9482 0.9658 0.9569 0.9898
0.0045 5.6667 24497 0.0764 0.9559 0.9598 0.9579 0.9905
0.0043 6.0 25938 0.0996 0.9510 0.9662 0.9585 0.9905
0.0037 6.3333 27379 0.0909 0.9539 0.9641 0.9590 0.9908
0.003 6.6667 28820 0.1010 0.9519 0.9639 0.9579 0.9905
0.003 7.0 30261 0.0944 0.9510 0.9632 0.9571 0.9905
0.0037 7.3333 31702 0.1041 0.9514 0.9642 0.9578 0.9903
0.0021 7.6667 33143 0.1048 0.9520 0.9658 0.9589 0.9907
0.0029 8.0 34584 0.1001 0.9526 0.9651 0.9588 0.9907
0.0019 8.3333 36025 0.1098 0.9525 0.9653 0.9588 0.9906
0.0019 8.6667 37466 0.1027 0.9538 0.9651 0.9594 0.9906
0.0019 9.0 38907 0.0990 0.9543 0.9653 0.9598 0.9908
0.0018 9.3333 40348 0.1086 0.9537 0.9655 0.9595 0.9907
0.0014 9.6667 41789 0.1090 0.9533 0.9658 0.9595 0.9907
0.0014 10.0 43230 0.1110 0.9525 0.9660 0.9592 0.9907

Framework versions

  • Transformers 4.41.1
  • Pytorch 2.3.0+cu121
  • Datasets 2.19.1
  • Tokenizers 0.19.1