whisper-tiny-kor_eng_tiny_pu_tx

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

• Loss: 2.5940
• Cer: 47.2433

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
4.5453	2.3256	100	3.3185	57.2330
2.7657	4.6512	200	2.4532	53.1262
1.7982	6.9767	300	2.0981	49.6635
1.0239	9.3023	400	2.1607	50.9805
0.52	11.6279	500	2.2176	50.9435
0.2505	13.9535	600	2.2748	50.1913
0.1193	16.2791	700	2.2759	48.0482
0.0671	18.6047	800	2.3094	48.9324
0.0419	20.9302	900	2.3421	48.7265
0.0282	23.2558	1000	2.3841	48.0509
0.0223	25.5814	1100	2.3495	46.9925
0.0167	27.9070	1200	2.4118	47.8107
0.0174	30.2326	1300	2.4403	48.3729
0.0111	32.5581	1400	2.4225	48.7028
0.0101	34.8837	1500	2.4477	47.3304
0.008	37.2093	1600	2.4688	48.0720
0.0066	39.5349	1700	2.4466	46.8843
0.0059	41.8605	1800	2.4788	48.5972
0.0044	44.1860	1900	2.5253	47.0268
0.0027	46.5116	2000	2.5079	45.9949
0.0022	48.8372	2100	2.5307	47.3436
0.0024	51.1628	2200	2.5216	47.3224
0.0017	53.4884	2300	2.5533	48.6500
0.001	55.8140	2400	2.5487	47.1958
0.0007	58.1395	2500	2.5486	46.4251
0.0006	60.4651	2600	2.5555	46.0318
0.0005	62.7907	2700	2.5608	45.9975
0.0005	65.1163	2800	2.5658	46.4013
0.0005	67.4419	2900	2.5694	46.2588
0.0004	69.7674	3000	2.5741	46.7840
0.0004	72.0930	3100	2.5770	46.9952
0.0004	74.4186	3200	2.5805	46.9186
0.0004	76.7442	3300	2.5832	46.9134
0.0004	79.0698	3400	2.5860	47.1509
0.0004	81.3953	3500	2.5885	46.9450
0.0004	83.7209	3600	2.5905	46.9292
0.0003	86.0465	3700	2.5918	47.1192
0.0003	88.3721	3800	2.5931	47.2881
0.0003	90.6977	3900	2.5936	47.3383
0.0003	93.0233	4000	2.5940	47.2433

Framework versions

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