Rhuax's picture
End of training
2ad65da
|
raw
history blame
2.31 kB
metadata
tags:
  - generated_from_trainer
datasets:
  - sms_spam
metrics:
  - accuracy
model-index:
  - name: MiniLMv2-L12-H384-distilled-finetuned-spam-detection
    results:
      - task:
          name: Text Classification
          type: text-classification
        dataset:
          name: sms_spam
          type: sms_spam
          args: plain_text
        metrics:
          - name: Accuracy
            type: accuracy
            value: 0.978494623655914

MiniLMv2-L12-H384-distilled-finetuned-spam-detection

This model is a fine-tuned version of nreimers/MiniLMv2-L12-H384-distilled-from-RoBERTa-Large on the sms_spam dataset. It achieves the following results on the evaluation set:

  • Loss: 0.4473
  • Accuracy: 0.9785

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: 0.0001
  • train_batch_size: 32
  • eval_batch_size: 32
  • seed: 33
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • num_epochs: 10
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss Accuracy
4.1186 1.0 18 3.4012 0.8351
2.9893 2.0 36 2.9206 0.8351
2.6718 3.0 54 2.8932 0.8351
2.5495 4.0 72 2.8916 0.8351
1.7213 5.0 90 0.6804 0.9821
0.7464 6.0 108 0.6017 0.9713
1.2052 7.0 126 0.3425 0.9857
0.438 8.0 144 0.2136 0.9857
0.2282 9.0 162 0.4539 0.9785
0.438 10.0 180 0.4473 0.9785

Framework versions

  • Transformers 4.17.0
  • Pytorch 1.10.2+cu113
  • Datasets 1.18.4
  • Tokenizers 0.12.1