whisper-tiny-kor_eng_tiny_na_na

This model was trained from scratch on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.7300
• Cer: 10.0074

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: 3e-05
• train_batch_size: 12
• eval_batch_size: 6
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• training_steps: 4000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Cer
3.148	0.1263	100	2.5953	39.9228
2.2793	0.2525	200	1.9118	30.5886
1.6733	0.3788	300	1.4350	25.9456
1.2947	0.5051	400	1.1628	20.4733
1.096	0.6313	500	1.0006	17.6024
0.9657	0.7576	600	0.9090	14.9178
0.8505	0.8838	700	0.8292	13.6419
0.8149	1.0101	800	0.7812	11.8369
0.5725	1.1364	900	0.7679	11.9624
0.5377	1.2626	1000	0.7441	11.7267
0.5448	1.3889	1100	0.7351	11.3596
0.5168	1.5152	1200	0.7160	11.2236
0.5314	1.6414	1300	0.7062	11.0735
0.5107	1.7677	1400	0.6988	10.8799
0.4962	1.8939	1500	0.6913	10.9093
0.4562	2.0202	1600	0.6914	10.6254
0.2799	2.1465	1700	0.7033	10.7638
0.2846	2.2727	1800	0.6973	10.4882
0.293	2.3990	1900	0.6964	10.2818
0.2966	2.5253	2000	0.6909	10.3885
0.2835	2.6515	2100	0.6866	10.1153
0.2925	2.7778	2200	0.6877	10.3041
0.2972	2.9040	2300	0.6823	10.1211
0.2574	3.0303	2400	0.7000	10.2466
0.1723	3.1566	2500	0.7070	10.0461
0.1517	3.2828	2600	0.7085	10.3193
0.17	3.4091	2700	0.7123	10.2138
0.1567	3.5354	2800	0.7131	10.3147
0.1511	3.6616	2900	0.7100	10.1376
0.1642	3.7879	3000	0.7103	10.2068
0.1569	3.9141	3100	0.7066	10.1645
0.1323	4.0404	3200	0.7186	9.7013
0.0872	4.1667	3300	0.7235	10.1305
0.0943	4.2929	3400	0.7315	10.0097
0.0958	4.4192	3500	0.7292	10.2689
0.0924	4.5455	3600	0.7315	9.8127
0.0836	4.6717	3700	0.7306	9.9652
0.0863	4.7980	3800	0.7308	10.1833
0.0957	4.9242	3900	0.7301	10.1716
0.0802	5.0505	4000	0.7300	10.0074

Framework versions

• Transformers 4.51.3
• Pytorch 2.6.0+cu126
• Datasets 3.5.0
• Tokenizers 0.21.1