whispherMusic

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

• Loss: 0.0963
• Rouge1: 90.9672
• Rouge2: 87.2313
• Rougel: 89.5998
• Rougelsum: 89.7339
• Gen Len: 61.55

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: 1e-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
• num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len
1.1983	1.0	1361	0.8877	47.1749	24.9442	38.1501	38.1934	56.0
0.8787	2.0	2722	0.6673	53.2613	32.6244	45.2405	45.3614	58.53
0.6983	3.0	4083	0.4976	57.5956	38.8147	51.0468	51.1725	59.97
0.5077	4.0	5444	0.3677	65.1283	49.4333	59.061	59.107	58.67
0.3955	5.0	6805	0.2650	70.453	58.4358	66.2936	66.5477	59.05
0.2846	6.0	8166	0.1987	77.3147	67.0836	73.4161	73.6763	59.26
0.21	7.0	9527	0.1489	84.5594	78.1538	82.1324	82.1614	60.14
0.1598	8.0	10888	0.1196	88.5138	83.886	86.8481	86.9753	61.14
0.1213	9.0	12249	0.1023	91.0285	87.2329	89.6909	89.7877	61.19
0.1051	10.0	13610	0.0963	90.9672	87.2313	89.5998	89.7339	61.55

Framework versions

• Transformers 4.31.0
• Pytorch 2.0.1+cu117
• Datasets 2.14.2
• Tokenizers 0.13.3