whisper_werbest_new_split

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

• Train Loss: 0.0590
• Train Accuracy: 0.0333
• Train Wermet: 13.3826
• Validation Loss: 0.4672
• Validation Accuracy: 0.0313
• Validation Wermet: 16.2097
• Epoch: 21

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:

• optimizer: {'name': 'AdamWeightDecay', 'learning_rate': 1e-05, 'decay': 0.0, 'beta_1': 0.9, 'beta_2': 0.999, 'epsilon': 1e-07, 'amsgrad': False, 'weight_decay_rate': 0.01}
• training_precision: float32

Training results

Train Loss	Train Accuracy	Train Wermet	Validation Loss	Validation Accuracy	Validation Wermet	Epoch
5.0901	0.0113	53.3790	4.4090	0.0122	42.3548	0
4.3135	0.0127	42.3551	3.9430	0.0149	37.1045	1
3.3458	0.0173	31.6069	2.3945	0.0222	25.5461	2
1.9669	0.0232	13.7935	1.4966	0.0261	6.9562	3
1.2830	0.0262	10.0196	1.1100	0.0279	9.5683	4
0.9517	0.0278	8.1513	0.9065	0.0289	7.8180	5
0.7555	0.0287	7.5457	0.7892	0.0295	5.1479	6
0.6204	0.0295	7.0748	0.7025	0.0299	6.9938	7
0.5202	0.0300	7.2085	0.6409	0.0303	7.6979	8
0.4418	0.0305	6.6665	0.5963	0.0305	4.9877	9
0.3773	0.0309	6.3833	0.5633	0.0307	5.6072	10
0.3239	0.0313	6.3658	0.5361	0.0308	9.7748	11
0.2784	0.0316	7.6413	0.5146	0.0310	8.5224	12
0.2390	0.0319	8.3862	0.5053	0.0310	8.1694	13
0.2049	0.0321	8.4188	0.4899	0.0311	9.4708	14
0.1749	0.0323	8.7733	0.4805	0.0312	8.5083	15
0.1480	0.0326	8.1859	0.4735	0.0312	16.2408	16
0.1242	0.0328	10.7089	0.4745	0.0312	6.8974	17
0.1042	0.0329	10.2003	0.4675	0.0313	9.7003	18
0.0862	0.0331	10.7710	0.4677	0.0313	6.6251	19
0.0708	0.0332	9.1255	0.4698	0.0313	13.2089	20
0.0590	0.0333	13.3826	0.4672	0.0313	16.2097	21

Framework versions

• Transformers 4.27.0.dev0
• TensorFlow 2.11.0
• Tokenizers 0.13.2