learning_rate: 2.8e-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	0.9909	95.7604	38.1377
0.9565	0.0492	3	0.9361	91.6129	36.1796
0.8142	0.0656	4	0.8502	87.6498	32.2233
0.7791	0.0820	5	0.7809	106.4516	43.1198
0.7181	0.0984	6	0.7391	103.5023	37.1520
0.7438	0.1148	7	0.7072	100.7373	36.4859
0.748	0.1311	8	0.7048	98.9862	43.3995
0.5673	0.1475	9	0.6841	92.8111	30.8912
0.5396	0.1639	10	0.6634	93.2719	32.7694
0.6983	0.1803	11	0.6445	83.4101	29.4125
0.6708	0.1967	12	0.6368	82.4885	27.9739
0.7027	0.2131	13	0.6216	94.1935	33.5553
0.6051	0.2295	14	0.6210	86.3594	31.8769
0.5511	0.2459	15	0.6240	83.2258	32.1433
0.5768	0.2623	16	0.6162	90.6912	38.3775
0.4678	0.2787	17	0.6087	89.2166	46.0903
0.4856	0.2951	18	0.6052	84.3318	34.3812
0.5423	0.3115	19	0.6066	77.6037	26.4953
0.456	0.3279	20	0.6049	82.3963	29.2127
0.533	0.3443	21	0.5943	80.6452	28.3735
0.5197	0.3607	22	0.5928	75.5760	25.5228
0.5618	0.3770	23	0.5947	74.1935	24.2707
0.5386	0.3934	24	0.5865	72.9954	23.7645
0.5499	0.4098	25	0.5800	80.3687	27.8007
0.4538	0.4262	26	0.5705	73.6406	23.8577
0.4069	0.4426	27	0.5721	73.8249	24.1774
0.4858	0.4590	28	0.5713	73.8249	24.3772
0.4643	0.4754	29	0.5685	74.1935	24.1508
0.4415	0.4918	30	0.5633	75.2995	24.0309
0.5562	0.5082	31	0.5584	73.0876	23.3515
0.507	0.5246	32	0.5559	72.8111	23.6180
0.4317	0.5410	33	0.5533	73.6406	23.9377
0.5814	0.5574	34	0.5479	74.4700	23.7512
0.4816	0.5738	35	0.5457	75.3917	23.9643
0.457	0.5902	36	0.5403	73.7327	23.7112
0.4626	0.6066	37	0.5332	73.2719	23.5913
0.533	0.6230	38	0.5310	71.0599	23.0318
0.4615	0.6393	39	0.5299	74.4700	26.9881
0.5341	0.6557	40	0.5330	71.4286	22.7255
0.4528	0.6721	41	0.5298	71.3364	22.5922
0.4833	0.6885	42	0.5273	70.6912	22.2725
0.4478	0.7049	43	0.5262	71.3364	22.7121
0.4476	0.7213	44	0.5228	71.0599	22.6855
0.4064	0.7377	45	0.5176	70.4147	22.5123
0.4493	0.7541	46	0.5136	69.3088	22.5922
0.4826	0.7705	47	0.5144	68.5714	22.8320
0.4807	0.7869	48	0.5173	67.9263	22.6589
0.3942	0.8033	49	0.5184	69.8618	24.3373
0.5007	0.8197	50	0.5167	69.4009	23.9110
0.4842	0.8361	51	0.5126	70.0461	24.2840
0.4101	0.8525	52	0.5076	70.4147	23.8311
0.5698	0.8689	53	0.5056	69.1244	22.3125
0.3463	0.8852	54	0.5047	68.7558	22.1926
0.4747	0.9016	55	0.5046	69.5853	22.1926
0.4461	0.9180	56	0.5068	73.6406	23.8977
0.3843	0.9344	57	0.5063	74.1935	24.8302
0.407	0.9508	58	0.5022	73.6406	25.8292
0.4396	0.9672	59	0.5011	70.0461	24.5637
0.4578	0.9836	60	0.5018	74.1935	25.8159
0.3383	1.0	61	0.5050	75.1152	25.3097
0.3131	1.0164	62	0.5031	72.0737	23.4847
0.2264	1.0328	63	0.5032	74.0092	24.6969
0.1684	1.0492	64	0.5132	73.9171	23.8577
0.2404	1.0656	65	0.5186	73.6406	23.5114
0.2013	1.0820	66	0.5178	72.4424	23.0851
0.2211	1.0984	67	0.5189	73.7327	24.1508
0.2675	1.1148	68	0.5178	66.3594	21.0337
0.2867	1.1311	69	0.5158	70.5991	23.1384
0.2904	1.1475	70	0.5118	70.6912	23.0585
0.236	1.1639	71	0.5138	69.4931	22.4058
0.1899	1.1803	72	0.5187	70.6912	22.5789
0.2274	1.1967	73	0.5192	70.8756	22.6322
0.2296	1.2131	74	0.5131	70.2304	22.2059
0.1956	1.2295	75	0.5080	70.5991	22.1260
0.2543	1.2459	76	0.5045	72.5346	22.4058
0.2275	1.2623	77	0.5031	69.6774	21.1536
0.2496	1.2787	78	0.4999	69.4009	20.8605
0.2103	1.2951	79	0.4978	67.5576	20.5541
0.1934	1.3115	80	0.4987	67.0968	20.4875
0.2214	1.3279	81	0.4968	66.9124	20.4343
0.2284	1.3443	82	0.4953	67.7419	20.5408
0.2301	1.3607	83	0.4923	67.6498	20.7140
0.2228	1.3770	84	0.4893	67.8341	20.7273
0.197	1.3934	85	0.4871	68.7558	20.7406
0.1991	1.4098	86	0.4864	70.4147	20.8472
0.2082	1.4262	87	0.4850	70.9677	21.1536
0.2074	1.4426	88	0.4841	68.2949	20.6074
0.2198	1.4590	89	0.4846	67.0046	20.7406
0.1936	1.4754	90	0.4843	67.2811	20.4742
0.2372	1.4918	91	0.4833	67.1889	20.4209
0.206	1.5082	92	0.4832	67.0046	20.7273
0.1344	1.5246	93	0.4824	66.7281	20.7007
0.2114	1.5410	94	0.4823	67.4654	20.8206
0.2009	1.5574	95	0.4814	67.3733	20.8072
0.2441	1.5738	96	0.4807	67.3733	20.9271
0.2333	1.5902	97	0.4807	67.4654	20.9671
0.2053	1.6066	98	0.4805	67.2811	20.9005
0.1697	1.6230	99	0.4799	66.9124	20.9005
0.1823	1.6393	100	0.4801	67.2811	20.9271
0.1679	1.6557	101	0.4800	67.1889	20.9138
0.1963	1.6721	102	0.4802	67.6498	21.0071
0.2069	1.6885	103	0.4800	67.3733	20.9671
0.2187	1.7049	104	0.4800	67.4654	20.8739
0.2565	1.7213	105	0.4805	67.2811	20.9005
0.177	1.7377	106	0.4806	66.9124	20.8739
0.1939	1.7541	107	0.4810	66.5438	20.7540
0.1748	1.7705	108	0.4812	67.4654	20.9405
0.18	1.7869	109	0.4813	66.9124	20.8339
0.1884	1.8033	110	0.4816	67.1889	20.8605
0.2282	1.8197	111	0.4812	66.5438	20.7140
0.1715	1.8361	112	0.4812	66.7281	20.9405
0.1678	1.8525	113	0.4806	66.2673	20.7673
0.1709	1.8689	114	0.4808	66.6359	20.7273
0.223	1.8852	115	0.4804	66.4516	20.7406
0.2245	1.9016	116	0.4806	66.4516	20.8739
0.2237	1.9180	117	0.4804	66.6359	20.7140
0.229	1.9344	118	0.4801	66.3594	20.7406
0.1952	1.9508	119	0.4802	66.2673	20.7806
0.1826	1.9672	120	0.4802	66.0829	20.7007
0.1967	1.9836	121	0.4803	66.0829	20.7007
0.1924	2.0	122	0.4801	66.4516	20.7806

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_2.8e-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_2.8e-5