whisper-tiny-kor_eng_tiny_oc_is

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

• Loss: 0.7291
• Cer: 9.6053

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
0.9204	8.3333	100	0.6321	12.2368
0.1223	16.6667	200	0.6041	12.5
0.0057	25.0	300	0.6492	9.5395
0.004	33.3333	400	0.6377	10.1974
0.0195	41.6667	500	0.6121	7.5
0.0099	50.0	600	0.5790	9.7368
0.001	58.3333	700	0.6290	10.3289
0.0002	66.6667	800	0.6388	10.3289
0.0001	75.0	900	0.6471	10.3289
0.0001	83.3333	1000	0.6540	10.3289
0.0001	91.6667	1100	0.6592	10.3289
0.0001	100.0	1200	0.6649	10.2632
0.0001	108.3333	1300	0.6700	10.3947
0.0001	116.6667	1400	0.6739	10.3289
0.0001	125.0	1500	0.6773	10.3289
0.0001	133.3333	1600	0.6809	10.3289
0.0	141.6667	1700	0.6841	10.3289
0.0	150.0	1800	0.6869	10.3289
0.0	158.3333	1900	0.6910	10.6579
0.0	166.6667	2000	0.6942	10.6579
0.0	175.0	2100	0.6970	10.6579
0.0	183.3333	2200	0.6986	10.6579
0.0	191.6667	2300	0.7015	10.2632
0.0	200.0	2400	0.7044	10.2632
0.0	208.3333	2500	0.7074	10.2632
0.0	216.6667	2600	0.7097	9.4737
0.0	225.0	2700	0.7120	9.4737
0.0	233.3333	2800	0.7148	9.4737
0.0	241.6667	2900	0.7164	9.6053
0.0	250.0	3000	0.7190	9.6053
0.0	258.3333	3100	0.7209	9.6053
0.0	266.6667	3200	0.7221	9.6053
0.0	275.0	3300	0.7228	9.6053
0.0	283.3333	3400	0.7246	9.6053
0.0	291.6667	3500	0.7257	9.6053
0.0	300.0	3600	0.7267	9.6053
0.0	308.3333	3700	0.7279	9.6053
0.0	316.6667	3800	0.7285	9.6053
0.0	325.0	3900	0.7289	9.6053
0.0	333.3333	4000	0.7291	9.6053

Framework versions

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