Hugging Face's logo

Rolo
/
distilhubert-finetuned-gtzan-4

Audio Classification
Transformers
PyTorch
hubert
Generated from Trainer
Eval Results
Inference Endpoints
Model card Files Community
distilhubert-finetuned-gtzan-4 / README.md
Rolo's picture
Rolo
End of training
67d2bc1
preview code
|
raw
history blame contribute delete
6.7 kB
metadata
license: apache-2.0
base_model: ntu-spml/distilhubert
tags:
  - generated_from_trainer
datasets:
  - marsyas/gtzan
metrics:
  - accuracy
model-index:
  - name: distilhubert-finetuned-gtzan-4
    results:
      - task:
          name: Audio Classification
          type: audio-classification
        dataset:
          name: GTZAN
          type: marsyas/gtzan
          config: default
          split: train
          args: default
        metrics:
          - name: Accuracy
            type: accuracy
            value: 0.82

distilhubert-finetuned-gtzan-4

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

  • Loss: 2.8452
  • Accuracy: 0.82

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: 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: 0.1
  • num_epochs: 80

Training results

Training Loss Epoch Step Validation Loss Accuracy
2.2676 1.0 225 2.2400 0.33
1.9086 2.0 450 1.8716 0.55
1.417 3.0 675 1.4094 0.69
1.0484 4.0 900 1.1065 0.77
1.0122 5.0 1125 0.9172 0.75
0.8937 6.0 1350 1.1343 0.66
0.2979 7.0 1575 0.6102 0.83
0.3413 8.0 1800 1.1212 0.71
1.459 9.0 2025 1.5433 0.75
0.5177 10.0 2250 1.3990 0.78
0.0494 11.0 2475 2.7712 0.78
0.792 12.0 2700 2.7047 0.81
0.0 13.0 2925 2.8097 0.82
0.0 14.0 3150 3.3873 0.79
0.0 15.0 3375 2.6185 0.81
0.0 16.0 3600 3.0773 0.81
0.0 17.0 3825 2.2380 0.84
0.0 18.0 4050 2.5949 0.79
0.0 19.0 4275 3.4890 0.75
0.0 20.0 4500 2.7776 0.82
0.0 21.0 4725 3.8952 0.77
0.0 22.0 4950 2.8020 0.8
0.0 23.0 5175 3.7968 0.72
0.0 24.0 5400 2.8630 0.81
0.0 25.0 5625 2.2026 0.84
0.0 26.0 5850 1.8612 0.87
0.0 27.0 6075 3.2650 0.78
0.0 28.0 6300 2.6739 0.82
0.0 29.0 6525 3.0167 0.8
0.0 30.0 6750 1.9316 0.84
0.0 31.0 6975 3.0458 0.8
0.0 32.0 7200 3.3569 0.77
0.0 33.0 7425 3.0012 0.79
0.0 34.0 7650 3.2477 0.79
0.0 35.0 7875 3.2145 0.79
0.0 36.0 8100 3.0645 0.79
0.0 37.0 8325 3.2974 0.77
0.0 38.0 8550 3.4422 0.77
0.0 39.0 8775 2.7268 0.8
0.0 40.0 9000 2.6908 0.8
0.0 41.0 9225 2.4034 0.82
0.0 42.0 9450 3.1446 0.79
0.0 43.0 9675 2.9127 0.8
0.0 44.0 9900 2.3812 0.81
0.0 45.0 10125 2.4215 0.81
0.0 46.0 10350 2.6125 0.82
0.7338 47.0 10575 2.5113 0.82
0.0 48.0 10800 2.9264 0.81
0.0 49.0 11025 2.7811 0.81
0.0 50.0 11250 2.6749 0.8
0.0 51.0 11475 3.2003 0.78
0.0 52.0 11700 3.2670 0.78
0.0 53.0 11925 3.4001 0.76
0.0 54.0 12150 2.8570 0.76
0.0 55.0 12375 2.1772 0.83
0.0 56.0 12600 2.7977 0.81
0.0 57.0 12825 2.9106 0.78
0.0 58.0 13050 2.8428 0.8
0.0 59.0 13275 2.7308 0.78
0.0 60.0 13500 2.8214 0.8
0.0 61.0 13725 2.8194 0.79
0.0 62.0 13950 1.7708 0.85
0.0 63.0 14175 2.6017 0.81
0.0 64.0 14400 2.7698 0.78
0.0 65.0 14625 2.8218 0.81
0.0 66.0 14850 2.8252 0.82
0.0 67.0 15075 2.9149 0.81
0.0 68.0 15300 2.8106 0.81
0.0 69.0 15525 2.8514 0.8
0.0 70.0 15750 2.6649 0.82
0.0 71.0 15975 2.5629 0.81
0.0 72.0 16200 2.8140 0.81
0.0 73.0 16425 2.8164 0.79
0.0 74.0 16650 2.7022 0.81
0.0 75.0 16875 2.7376 0.81
0.0 76.0 17100 2.6498 0.8
0.0 77.0 17325 2.7363 0.81
0.0 78.0 17550 2.8057 0.81
0.0 79.0 17775 2.8526 0.81
0.0 80.0 18000 2.8452 0.82

Framework versions

  • Transformers 4.33.0.dev0
  • Pytorch 2.0.1+cu117
  • Datasets 2.14.4
  • Tokenizers 0.13.3