whisper-tiny-kor_eng_tiny_ps_is

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

• Loss: 0.9362
• Cer: 18.5185

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
0.913	7.1429	100	0.8864	14.2119
0.1679	14.2857	200	0.7422	13.1496
0.0061	21.4286	300	0.7671	14.1832
0.0015	28.5714	400	0.7874	13.2644
0.0008	35.7143	500	0.8058	13.1209
0.0005	42.8571	600	0.8204	13.6951
0.0003	50.0	700	0.8321	13.7238
0.0002	57.1429	800	0.8413	13.5515
0.0002	64.2857	900	0.8489	13.6664
0.0002	71.4286	1000	0.8560	14.5564
0.0001	78.5714	1100	0.8625	13.2931
0.0001	85.7143	1200	0.8678	12.7476
0.0001	92.8571	1300	0.8733	13.1209
0.0001	100.0	1400	0.8776	14.4703
0.0001	107.1429	1500	0.8821	14.7574
0.0001	114.2857	1600	0.8861	14.5851
0.0001	121.4286	1700	0.8901	14.2980
0.0001	128.5714	1800	0.8938	16.3365
0.0001	135.7143	1900	0.8977	16.5088
0.0001	142.8571	2000	0.9011	16.4800
0.0001	150.0	2100	0.9077	17.1691
0.0	157.1429	2200	0.9108	17.1404
0.0	164.2857	2300	0.9134	17.1404
0.0	171.4286	2400	0.9155	17.1404
0.0	178.5714	2500	0.9173	17.1117
0.0	185.7143	2600	0.9202	18.4037
0.0	192.8571	2700	0.9221	17.9156
0.0	200.0	2800	0.9234	17.9443
0.0	207.1429	2900	0.9256	17.9443
0.0	214.2857	3000	0.9272	17.9443
0.0	221.4286	3100	0.9293	18.4611
0.0	228.5714	3200	0.9305	18.4611
0.0	235.7143	3300	0.9316	18.4611
0.0	242.8571	3400	0.9329	18.5185
0.0	250.0	3500	0.9344	18.5185
0.0	257.1429	3600	0.9348	18.5185
0.0	264.2857	3700	0.9354	18.4898
0.0	271.4286	3800	0.9357	18.5185
0.0	278.5714	3900	0.9356	18.5185
0.0	285.7143	4000	0.9362	18.5185

Framework versions

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