deopoler's picture
End of training
3e1eb60 verified
metadata
license: apache-2.0
base_model: ntu-spml/distilhubert
tags:
  - generated_from_trainer
datasets:
  - mechanicalkeystrokes
metrics:
  - accuracy
model-index:
  - name: distilhubert-finetuned-mechanicalkeystrokes
    results:
      - task:
          name: Audio Classification
          type: audio-classification
        dataset:
          name: MechanicalKeystrokes
          type: mechanicalkeystrokes
          config: default
          split: train
          args: default
        metrics:
          - name: Accuracy
            type: accuracy
            value: 0.9666666666666667

distilhubert-finetuned-mechanicalkeystrokes

This model is a fine-tuned version of ntu-spml/distilhubert on the MechanicalKeystrokes dataset. It achieves the following results on the evaluation set:

  • Loss: 0.1155
  • Accuracy: 0.9667

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-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
  • lr_scheduler_warmup_ratio: 0.1
  • num_epochs: 50
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss Accuracy
2.2923 1.0 30 2.3063 0.0333
2.2788 2.0 60 2.2863 0.1333
2.1377 3.0 90 2.1583 0.35
1.8769 4.0 120 1.8170 0.6167
1.5083 5.0 150 1.4282 0.7833
1.0911 6.0 180 1.0241 0.9333
0.7235 7.0 210 0.7131 0.9667
0.4517 8.0 240 0.4509 0.9667
0.2529 9.0 270 0.2682 0.9833
0.14 10.0 300 0.1799 0.9833
0.0791 11.0 330 0.1155 0.9667

Framework versions

  • Transformers 4.40.1
  • Pytorch 2.2.1
  • Datasets 2.19.0
  • Tokenizers 0.19.1