metadata

license: apache-2.0
tags:
  - generated_from_trainer
datasets:
  - marsyas/gtzan
metrics:
  - accuracy
model-index:
  - name: distilhubert-finetuned-gtzan
    results: []

distilhubert-finetuned-gtzan

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: 0.7667
Accuracy: 0.88

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
distributed_type: multi-GPU
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
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
2.1524	1.0	225	2.0279	0.45
1.2284	2.0	450	1.3462	0.62
1.014	3.0	675	0.9385	0.71
1.2816	4.0	900	0.8428	0.75
0.3312	5.0	1125	0.5206	0.83
0.7004	6.0	1350	0.9608	0.76
0.0515	7.0	1575	0.6214	0.85
0.0114	8.0	1800	0.7193	0.83
0.0032	9.0	2025	0.7997	0.86
0.0021	10.0	2250	1.0831	0.81
0.0059	11.0	2475	0.9561	0.83
0.0011	12.0	2700	0.7667	0.88
0.0008	13.0	2925	0.8389	0.87
0.0007	14.0	3150	0.8570	0.87
0.0006	15.0	3375	0.8778	0.86
0.0005	16.0	3600	0.9170	0.87
0.0004	17.0	3825	0.9422	0.87
0.0003	18.0	4050	0.9408	0.87
0.0005	19.0	4275	0.8940	0.87
0.0003	20.0	4500	0.9724	0.86
0.0003	21.0	4725	0.8904	0.85
0.0002	22.0	4950	0.9573	0.86
0.0002	23.0	5175	0.9292	0.87
0.0002	24.0	5400	0.9209	0.86
0.0002	25.0	5625	0.9184	0.86
0.0002	26.0	5850	0.9005	0.85
0.0002	27.0	6075	0.9656	0.86
0.0002	28.0	6300	0.9685	0.86
0.0002	29.0	6525	0.9810	0.86
0.0002	30.0	6750	0.9860	0.86

Framework versions

Transformers 4.29.2
Pytorch 1.13.1+cu117
Datasets 2.12.0
Tokenizers 0.13.3