learning_rate: 0.001
train_batch_size: 32
eval_batch_size: 32
seed: 42
optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED 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: 0.1
num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Wer	Cer
2.1077	0.1433	100	2.0875	1.1545	0.6442
1.3038	0.2865	200	1.2774	0.9410	0.4603
1.1622	0.4298	300	1.1382	0.9318	0.4292
1.0598	0.5731	400	1.0422	0.8139	0.3616
0.9698	0.7163	500	0.9831	0.8650	0.4124
1.0004	0.8596	600	0.9329	0.8296	0.3977
0.9246	1.0029	700	0.8990	0.7937	0.3763
0.8547	1.1461	800	0.8675	0.8520	0.4602
0.9163	1.2894	900	0.8364	0.7504	0.3477
0.8319	1.4327	1000	0.8205	0.7264	0.3456
0.8336	1.5759	1100	0.7872	0.7060	0.3151
0.7520	1.7192	1200	0.7683	0.7115	0.3259
0.7394	1.8625	1300	0.7534	0.6717	0.2963
0.6487	2.0057	1400	0.7362	0.6918	0.3321
0.6508	2.1490	1500	0.7349	0.6851	0.3106
0.6892	2.2923	1600	0.7207	0.7103	0.3391
0.6132	2.4355	1700	0.7182	0.6812	0.3050
0.6361	2.5788	1800	0.7013	0.7311	0.3462
0.6132	2.7221	1900	0.6935	0.6683	0.3133
0.6241	2.8653	2000	0.6861	0.7198	0.3492
0.5422	3.0086	2100	0.6782	0.6610	0.3011
0.5390	3.1519	2200	0.6831	0.6422	0.2983
0.5884	3.2951	2300	0.6734	0.6759	0.3239
0.5608	3.4384	2400	0.6720	0.6720	0.3242
0.5315	3.5817	2500	0.6606	0.6628	0.2952
0.4802	3.7249	2600	0.6539	0.6310	0.2849
0.5657	3.8682	2700	0.6458	0.6120	0.2736
0.4700	4.0115	2800	0.6510	0.6163	0.2805
0.4061	4.1547	2900	0.6600	0.6365	0.2918
0.4719	4.2980	3000	0.6517	0.6253	0.2820
0.4679	4.4413	3100	0.6493	0.5976	0.2680
0.4682	4.5845	3200	0.6440	0.6113	0.2746
0.4154	4.7278	3300	0.6364	0.6146	0.2859
0.5108	4.8711	3400	0.6324	0.5990	0.2680
0.3766	5.0143	3500	0.6392	0.5850	0.2659
0.3648	5.1576	3600	0.6478	0.6036	0.2768
0.3899	5.3009	3700	0.6429	0.6271	0.2914
0.4160	5.4441	3800	0.6422	0.6005	0.2725
0.4492	5.5874	3900	0.6426	0.6126	0.2843
0.3872	5.7307	4000	0.6377	0.5907	0.2726
0.4025	5.8739	4100	0.6374	0.5883	0.2673
0.2973	6.0172	4200	0.6476	0.5874	0.2697
0.2923	6.1605	4300	0.6602	0.6035	0.2777
0.2842	6.3037	4400	0.6654	0.5988	0.2764
0.3156	6.4470	4500	0.6530	0.6013	0.2767
0.3275	6.5903	4600	0.6513	0.6081	0.2814
0.2956	6.7335	4700	0.6502	0.5862	0.2713
0.2636	6.8768	4800	0.6507	0.5942	0.2769
0.2363	7.0201	4900	0.6660	0.5952	0.2780
0.2353	7.1633	5000	0.6786	0.6004	0.2811
0.2229	7.3066	5100	0.6801	0.6075	0.2845
0.2273	7.4499	5200	0.6838	0.5753	0.2628
0.2257	7.5931	5300	0.6772	0.5986	0.2773
0.2647	7.7364	5400	0.6763	0.5918	0.2764
0.2646	7.8797	5500	0.6764	0.5934	0.2731
0.1760	8.0229	5600	0.7018	0.5955	0.2724
0.1940	8.1662	5700	0.7091	0.5901	0.2689
0.1803	8.3095	5800	0.7106	0.6044	0.2804
0.1810	8.4527	5900	0.7054	0.5898	0.2679
0.1964	8.5960	6000	0.7074	0.6034	0.2812
0.2025	8.7393	6100	0.7128	0.5893	0.2706
0.1862	8.8825	6200	0.7079	0.5946	0.2733
0.1495	9.0258	6300	0.7216	0.5939	0.2731
0.1288	9.1691	6400	0.7314	0.5917	0.2742
0.1245	9.3123	6500	0.7353	0.5918	0.2757
0.1432	9.4556	6600	0.7376	0.5950	0.2741
0.1398	9.5989	6700	0.7387	0.5977	0.2800
0.1250	9.7421	6800	0.7390	0.6063	0.2859
0.1169	9.8854	6900	0.7369	0.6002	0.2809

Framework versions

PEFT 0.18.1
Transformers 5.3.0.dev0
Pytorch 2.10.0+cu128
Datasets 4.0.0
Tokenizers 0.22.2

sudoping01/karato-whisper-latin

karato-whisper-latin

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

sudoping01/
karato-whisper-latin