whisper-base-finetuned-gtzan

This model is a fine-tuned version of openai/whisper-base on the GTZAN dataset. It achieves the following results on the evaluation set:

• Loss: 3.8944
• Accuracy: 0.62

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: 30

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
2.3577	1.0	200	1.9551	0.35
2.0492	2.0	400	2.0333	0.27
2.0331	3.0	600	1.9196	0.3
1.3732	4.0	800	1.6705	0.34
1.7021	5.0	1000	1.7006	0.335
1.907	6.0	1200	1.7489	0.36
1.611	7.0	1400	1.5347	0.45
1.1989	8.0	1600	1.4835	0.465
2.0049	9.0	1800	1.3681	0.525
0.9562	10.0	2000	1.4732	0.49
0.4145	11.0	2200	1.2645	0.555
1.5859	12.0	2400	1.3992	0.51
1.5115	13.0	2600	1.2638	0.545
0.9777	14.0	2800	1.4003	0.57
0.831	15.0	3000	1.3377	0.575
1.3201	16.0	3200	1.5033	0.575
1.1711	17.0	3400	1.5239	0.555
0.4201	18.0	3600	1.6902	0.555
0.346	19.0	3800	1.9733	0.525
0.5619	20.0	4000	2.1321	0.555
0.645	21.0	4200	2.1219	0.625
0.2672	22.0	4400	2.2037	0.555
0.2826	23.0	4600	2.7297	0.565
0.4265	24.0	4800	3.3848	0.5
0.0319	25.0	5000	3.5627	0.59
0.0024	26.0	5200	3.7420	0.6
0.0332	27.0	5400	3.7159	0.63
0.0009	28.0	5600	3.8011	0.635
0.0001	29.0	5800	3.8852	0.615
0.0001	30.0	6000	3.8944	0.62

Framework versions

• Transformers 4.32.1
• Pytorch 2.0.1+cu118
• Datasets 2.14.4
• Tokenizers 0.13.3