learning_rate: 2.2e-05
train_batch_size: 32
eval_batch_size: 8
seed: 42
optimizer: Use adamw_torch_fused with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
lr_scheduler_type: cosine
lr_scheduler_warmup_ratio: 0.1
num_epochs: 2

Training results

Training Loss	Epoch	Step	Validation Loss	Wer	Cer
1.0467	0.0164	1	1.2545	98.1567	40.8419
1.1366	0.0328	2	0.9969	100.4608	40.0559
0.9618	0.0492	3	0.9433	92.3502	36.4593
0.8201	0.0656	4	0.8657	89.1244	33.1024
0.7934	0.0820	5	0.8056	89.1244	30.9311
0.7371	0.0984	6	0.7576	92.7189	33.3156
0.7617	0.1148	7	0.7260	90.9677	34.2614
0.7717	0.1311	8	0.7016	109.8618	41.5212
0.5622	0.1475	9	0.6797	98.1567	37.2452
0.5371	0.1639	10	0.6584	98.1567	32.5430
0.6886	0.1803	11	0.6384	96.4977	32.8227
0.6591	0.1967	12	0.6277	84.8848	28.1870
0.6858	0.2131	13	0.6189	80.3687	26.1489
0.5996	0.2295	14	0.6120	77.7880	25.3896
0.545	0.2459	15	0.6140	89.0323	31.3974
0.5714	0.2623	16	0.6073	81.8433	31.2375
0.4672	0.2787	17	0.6029	88.9401	45.7307
0.4809	0.2951	18	0.5958	85.7143	35.6734
0.5443	0.3115	19	0.5939	88.9401	29.7056
0.445	0.3279	20	0.5891	76.5899	24.4305
0.5222	0.3443	21	0.5787	81.7512	27.1347
0.5174	0.3607	22	0.5783	77.1429	28.8797
0.5426	0.3770	23	0.5919	74.3779	24.7103
0.533	0.3934	24	0.5825	72.8111	24.3106
0.5415	0.4098	25	0.5680	72.0737	23.4448
0.4445	0.4262	26	0.5652	82.1198	27.7741
0.4043	0.4426	27	0.5666	79.8157	27.1746
0.4958	0.4590	28	0.5567	70.9677	22.9119
0.4539	0.4754	29	0.5622	71.3364	23.6180
0.4294	0.4918	30	0.5668	70.7834	23.4581
0.5555	0.5082	31	0.5568	71.3364	23.1118
0.5043	0.5246	32	0.5463	72.5346	23.1118
0.4138	0.5410	33	0.5447	72.1659	23.1118
0.5678	0.5574	34	0.5419	75.1152	23.4315
0.4812	0.5738	35	0.5360	73.5484	23.2050
0.4521	0.5902	36	0.5327	72.1659	22.8986
0.4489	0.6066	37	0.5300	71.4286	22.8720
0.5172	0.6230	38	0.5264	69.9539	22.3924
0.458	0.6393	39	0.5194	73.1797	25.1632
0.525	0.6557	40	0.5153	69.3088	21.9262
0.4417	0.6721	41	0.5105	68.9401	22.0328
0.471	0.6885	42	0.5102	69.4009	22.2592
0.4417	0.7049	43	0.5097	69.4009	21.8596
0.4368	0.7213	44	0.5084	70.0461	21.9395
0.3992	0.7377	45	0.5051	69.5853	22.1527
0.432	0.7541	46	0.5032	69.8618	22.2059
0.4762	0.7705	47	0.5038	68.5714	22.2459
0.4783	0.7869	48	0.5057	68.0184	22.0328
0.3978	0.8033	49	0.5060	68.6636	22.0061
0.5011	0.8197	50	0.5050	68.4793	22.0994
0.4803	0.8361	51	0.5031	68.6636	22.2592
0.399	0.8525	52	0.5008	69.1244	21.9928
0.5721	0.8689	53	0.5000	69.5853	22.3791
0.3436	0.8852	54	0.4980	68.9401	22.0861
0.4632	0.9016	55	0.4961	69.1244	22.1260
0.443	0.9180	56	0.4964	72.6267	23.6046
0.3709	0.9344	57	0.4975	73.8249	23.8178
0.4071	0.9508	58	0.4942	74.8387	24.7502
0.4331	0.9672	59	0.4905	73.9171	24.4971
0.4466	0.9836	60	0.4898	68.4793	21.6731
0.3378	1.0	61	0.4930	81.5668	26.1889
0.3231	1.0164	62	0.4928	78.2488	24.5904
0.2373	1.0328	63	0.4911	70.9677	23.0984
0.1737	1.0492	64	0.4970	70.3226	23.2849
0.2477	1.0656	65	0.5023	68.5714	21.8330
0.2072	1.0820	66	0.5022	67.4654	21.4466
0.2245	1.0984	67	0.5002	71.2442	22.6855
0.2653	1.1148	68	0.4984	65.9908	20.5808
0.2913	1.1311	69	0.4974	68.2949	21.1536
0.3033	1.1475	70	0.4950	69.6774	21.4733
0.2467	1.1639	71	0.4942	71.3364	22.7654
0.1891	1.1803	72	0.4967	66.4516	21.3001
0.225	1.1967	73	0.4992	67.4654	21.1936
0.2322	1.2131	74	0.4981	66.8203	20.8872
0.1976	1.2295	75	0.4960	71.8894	22.6056
0.2555	1.2459	76	0.4942	71.5207	22.5123
0.2383	1.2623	77	0.4917	68.4793	20.7273
0.2498	1.2787	78	0.4891	68.3871	20.5275
0.2148	1.2951	79	0.4879	69.3088	20.7007
0.1962	1.3115	80	0.4878	68.5714	20.4076
0.2198	1.3279	81	0.4868	68.0184	20.3277
0.2388	1.3443	82	0.4868	68.2949	20.3543
0.2295	1.3607	83	0.4852	68.7558	20.5142
0.2234	1.3770	84	0.4834	69.4009	20.6740
0.2089	1.3934	85	0.4819	69.2166	20.6208
0.2101	1.4098	86	0.4806	72.9032	22.1660
0.2202	1.4262	87	0.4794	73.0876	22.2859
0.213	1.4426	88	0.4787	72.3502	22.1793
0.2289	1.4590	89	0.4781	67.6498	20.5275
0.2059	1.4754	90	0.4778	67.7419	20.4209
0.2496	1.4918	91	0.4764	67.2811	20.2611
0.2136	1.5082	92	0.4752	67.7419	20.4343
0.1397	1.5246	93	0.4749	67.2811	20.3543
0.2214	1.5410	94	0.4744	67.5576	20.3277
0.2123	1.5574	95	0.4733	67.4654	20.4609
0.2489	1.5738	96	0.4730	67.8341	20.3144
0.2405	1.5902	97	0.4729	66.9124	20.0879
0.2122	1.6066	98	0.4724	66.7281	20.2344
0.1745	1.6230	99	0.4720	67.0046	20.1012
0.1881	1.6393	100	0.4719	66.7281	20.0613
0.1737	1.6557	101	0.4722	66.9124	20.1945
0.1997	1.6721	102	0.4725	67.8341	20.3277
0.2096	1.6885	103	0.4726	66.8203	20.3410
0.2266	1.7049	104	0.4729	67.0046	20.2478
0.2651	1.7213	105	0.4726	66.1751	20.1279
0.1875	1.7377	106	0.4734	66.3594	20.2877
0.2011	1.7541	107	0.4739	66.5438	20.2478
0.1769	1.7705	108	0.4744	66.3594	20.1412
0.1861	1.7869	109	0.4743	66.3594	20.0346
0.1993	1.8033	110	0.4751	66.3594	20.0346
0.2387	1.8197	111	0.4747	67.0046	20.2344
0.1803	1.8361	112	0.4750	66.2673	20.0213
0.1784	1.8525	113	0.4746	66.8203	20.1146
0.1794	1.8689	114	0.4744	66.1751	20.0080
0.2426	1.8852	115	0.4743	66.7281	20.0746
0.2283	1.9016	116	0.4743	66.7281	20.1146
0.2412	1.9180	117	0.4741	66.6359	20.0879
0.2372	1.9344	118	0.4738	66.6359	20.0746
0.2024	1.9508	119	0.4736	66.8203	20.1279
0.195	1.9672	120	0.4739	66.9124	20.1545
0.2083	1.9836	121	0.4738	66.9124	20.1945
0.1995	2.0	122	0.4738	66.4516	20.0613

Framework versions

Transformers 4.57.3
Pytorch 2.9.1+cu128
Datasets 4.4.1
Tokenizers 0.22.1

nkkbr/whisper-large-v3-zatoichi-ja-JDG_ver_20260212_lr_2.2e-5

Whisper Large v3 - Japanese Zatoichi ASR

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

nkkbr/
whisper-large-v3-zatoichi-ja-JDG_ver_20260212_lr_2.2e-5