whisper-tiny-kor_eng_tiny_ps_op

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

• Loss: 0.6630
• Cer: 8.4739

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
0.9705	0.2632	100	0.8838	17.0641
0.7347	0.5263	200	0.7395	14.0577
0.6634	0.7895	300	0.6340	12.3528
0.5271	1.0526	400	0.5735	10.2134
0.3478	1.3158	500	0.5666	10.3097
0.3566	1.5789	600	0.5474	9.9771
0.3506	1.8421	700	0.5242	10.0734
0.2574	2.1053	800	0.5379	10.3061
0.1485	2.3684	900	0.5378	8.8611
0.1527	2.6316	1000	0.5364	8.6120
0.1659	2.8947	1100	0.5363	9.1737
0.092	3.1579	1200	0.5624	9.3046
0.0623	3.4211	1300	0.5694	9.1573
0.0702	3.6842	1400	0.5698	10.5533
0.0671	3.9474	1500	0.5669	10.1443
0.0361	4.2105	1600	0.5892	9.0465
0.0302	4.4737	1700	0.6030	9.1791
0.0296	4.7368	1800	0.5995	9.2137
0.0282	5.0	1900	0.6058	9.2191
0.0117	5.2632	2000	0.6340	10.1207
0.0135	5.5263	2100	0.6179	8.6866
0.0136	5.7895	2200	0.6205	8.7120
0.0117	6.0526	2300	0.6169	8.6538
0.0058	6.3158	2400	0.6275	8.6175
0.0055	6.5789	2500	0.6343	8.7120
0.0069	6.8421	2600	0.6419	8.8974
0.0043	7.1053	2700	0.6399	9.8880
0.0039	7.3684	2800	0.6410	8.5575
0.0028	7.6316	2900	0.6496	8.4194
0.0027	7.8947	3000	0.6514	8.5339
0.0021	8.1579	3100	0.6517	9.0156
0.0018	8.4211	3200	0.6555	8.6102
0.0019	8.6842	3300	0.6566	8.5430
0.0019	8.9474	3400	0.6577	8.4921
0.0019	9.2105	3500	0.6581	8.5611
0.0013	9.4737	3600	0.6612	8.5611
0.0012	9.7368	3700	0.6622	8.4921
0.0014	10.0	3800	0.6626	8.4412
0.0016	10.2632	3900	0.6630	8.4575
0.0011	10.5263	4000	0.6630	8.4739

Framework versions

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