metadata

license: apache-2.0
base_model: ntu-spml/distilhubert
tags:
  - generated_from_trainer
datasets:
  - marsyas/gtzan
metrics:
  - accuracy
model-index:
  - name: distilhubert-finetuned-gtzan4
    results:
      - task:
          name: Audio Classification
          type: audio-classification
        dataset:
          name: GTZAN
          type: marsyas/gtzan
          config: all
          split: train
          args: all
        metrics:
          - name: Accuracy
            type: accuracy
            value: 0.78

distilhubert-finetuned-gtzan4

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: 1.0945
Accuracy: 0.78

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: 6
eval_batch_size: 6
seed: 42
gradient_accumulation_steps: 32
total_train_batch_size: 192
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
lr_scheduler_warmup_ratio: 0.1
num_epochs: 30

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	0.85	4	2.2991	0.06
2.2997	1.92	9	2.2668	0.28
2.2819	2.99	14	2.1877	0.33
2.2336	3.84	18	2.1023	0.47
2.1493	4.91	23	1.9895	0.52
2.0571	5.97	28	1.8745	0.51
1.9341	6.83	32	1.7838	0.57
1.8274	7.89	37	1.6784	0.64
1.724	8.96	42	1.5859	0.66
1.6407	9.81	46	1.5234	0.66
1.5593	10.88	51	1.4508	0.7
1.4735	11.95	56	1.3982	0.69
1.4185	12.8	60	1.3501	0.72
1.3613	13.87	65	1.3131	0.74
1.3099	14.93	70	1.2742	0.72
1.2762	16.0	75	1.2485	0.73
1.2762	16.85	79	1.2102	0.74
1.2379	17.92	84	1.1931	0.75
1.193	18.99	89	1.1647	0.75
1.1863	19.84	93	1.1488	0.77
1.1435	20.91	98	1.1349	0.78
1.1424	21.97	103	1.1166	0.79
1.0961	22.83	107	1.1025	0.78
1.0887	23.89	112	1.0993	0.78
1.0977	24.96	117	1.0952	0.78
1.0661	25.6	120	1.0945	0.78

Framework versions

Transformers 4.32.0.dev0
Pytorch 1.13.1+cu117
Datasets 2.14.4
Tokenizers 0.13.3