learning_rate: 1.4e-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	0.9981	100.8295	40.1226
0.9628	0.0492	3	0.9694	94.1935	37.4983
0.8432	0.0656	4	0.9222	89.9539	35.2604
0.8449	0.0820	5	0.8580	88.9401	33.0358
0.7821	0.0984	6	0.8142	90.8756	32.3565
0.8153	0.1148	7	0.7644	90.7834	36.8856
0.8144	0.1311	8	0.7391	93.4562	38.6306
0.6041	0.1475	9	0.7100	95.9447	37.8447
0.564	0.1639	10	0.6911	100.9217	36.4859
0.7197	0.1803	11	0.6618	100.0	37.7514
0.6799	0.1967	12	0.6438	107.0046	37.1920
0.7038	0.2131	13	0.6333	90.1382	31.2242
0.6186	0.2295	14	0.6190	83.1336	27.7608
0.5577	0.2459	15	0.6090	84.2396	27.7608
0.5531	0.2623	16	0.6070	76.7742	24.5904
0.4753	0.2787	17	0.6016	82.5806	26.7084
0.4771	0.2951	18	0.5959	83.5945	30.4249
0.5545	0.3115	19	0.5877	79.3548	25.4562
0.4436	0.3279	20	0.5851	82.3963	28.4268
0.5181	0.3443	21	0.5785	81.6590	27.9606
0.5177	0.3607	22	0.5716	78.0645	26.6818
0.5366	0.3770	23	0.5796	77.4194	28.8531
0.5237	0.3934	24	0.5762	74.0092	26.2155
0.5373	0.4098	25	0.5669	72.3502	23.6180
0.4487	0.4262	26	0.5599	77.3272	24.7236
0.395	0.4426	27	0.5577	73.3641	23.3249
0.4817	0.4590	28	0.5542	78.8940	24.9900
0.4511	0.4754	29	0.5588	78.3410	24.8435
0.4358	0.4918	30	0.5522	72.7189	23.4448
0.5482	0.5082	31	0.5481	72.4424	23.5380
0.5053	0.5246	32	0.5495	71.7051	23.1917
0.4177	0.5410	33	0.5474	72.8111	23.1384
0.5667	0.5574	34	0.5406	73.4562	23.0318
0.4687	0.5738	35	0.5354	73.7327	23.1917
0.4541	0.5902	36	0.5319	80.4608	25.7360
0.4454	0.6066	37	0.5270	79.5392	25.4829
0.511	0.6230	38	0.5208	73.4562	22.6056
0.4703	0.6393	39	0.5167	71.5207	22.4191
0.5256	0.6557	40	0.5173	70.5069	21.9662
0.4419	0.6721	41	0.5152	71.8894	21.9928
0.4823	0.6885	42	0.5109	73.5484	24.5238
0.4552	0.7049	43	0.5058	74.1935	24.9234
0.4432	0.7213	44	0.5016	70.0461	22.2326
0.3936	0.7377	45	0.4983	69.4931	21.5799
0.4348	0.7541	46	0.4972	71.1521	21.9395
0.4879	0.7705	47	0.4965	70.5991	22.2193
0.4758	0.7869	48	0.4964	70.1382	22.0594
0.3998	0.8033	49	0.4966	69.6774	22.2059
0.5079	0.8197	50	0.4973	70.1382	22.0328
0.4743	0.8361	51	0.4961	70.5991	22.0994
0.3918	0.8525	52	0.4937	69.4931	21.6465
0.5702	0.8689	53	0.4924	69.4931	21.7131
0.345	0.8852	54	0.4903	73.3641	23.2849
0.4643	0.9016	55	0.4890	73.1797	23.2183
0.4393	0.9180	56	0.4905	73.0876	23.2716
0.3611	0.9344	57	0.4934	72.4424	23.2450
0.4097	0.9508	58	0.4924	71.7972	23.6180
0.4346	0.9672	59	0.4885	72.6267	23.6446
0.4457	0.9836	60	0.4852	72.9032	24.6170
0.3451	1.0	61	0.4849	72.6267	23.4448
0.3444	1.0164	62	0.4854	73.2719	23.0718
0.2574	1.0328	63	0.4859	68.3871	21.4866
0.1913	1.0492	64	0.4868	69.0323	21.6065
0.2677	1.0656	65	0.4882	67.1889	21.7397
0.2288	1.0820	66	0.4890	68.1106	21.1003
0.2423	1.0984	67	0.4894	67.8341	21.0337
0.2817	1.1148	68	0.4882	67.5576	21.0603
0.3196	1.1311	69	0.4873	66.5438	20.5275
0.327	1.1475	70	0.4863	66.3594	20.5808
0.2701	1.1639	71	0.4863	67.5576	20.8072
0.2099	1.1803	72	0.4872	67.4654	20.9405
0.2558	1.1967	73	0.4882	67.7419	20.9405
0.2471	1.2131	74	0.4881	67.7419	20.9271
0.2207	1.2295	75	0.4878	67.6498	20.9138
0.2822	1.2459	76	0.4870	68.5714	20.9804
0.2616	1.2623	77	0.4864	68.7558	20.9271
0.278	1.2787	78	0.4855	69.1244	21.0470
0.2461	1.2951	79	0.4847	72.0737	21.9928
0.2284	1.3115	80	0.4844	71.6129	21.7264
0.2435	1.3279	81	0.4837	71.5207	21.6598
0.2686	1.3443	82	0.4833	67.3733	19.9547
0.2648	1.3607	83	0.4824	68.0184	20.1279
0.2469	1.3770	84	0.4819	68.0184	20.2611
0.2287	1.3934	85	0.4810	68.2949	20.0480
0.2344	1.4098	86	0.4809	67.7419	19.8748
0.2436	1.4262	87	0.4802	68.1106	19.8481
0.2397	1.4426	88	0.4801	67.0968	19.7283
0.2616	1.4590	89	0.4796	67.4654	19.8348
0.23	1.4754	90	0.4794	66.8203	19.8348
0.2726	1.4918	91	0.4781	67.0046	19.9281
0.2342	1.5082	92	0.4775	67.2811	19.9947
0.1631	1.5246	93	0.4766	67.4654	19.9147
0.2532	1.5410	94	0.4767	66.6359	19.8881
0.2411	1.5574	95	0.4757	66.3594	19.7016
0.2854	1.5738	96	0.4752	66.1751	19.7283
0.2687	1.5902	97	0.4742	66.5438	19.7549
0.2404	1.6066	98	0.4738	66.1751	19.7815
0.1993	1.6230	99	0.4733	66.3594	19.7682
0.213	1.6393	100	0.4729	66.4516	19.8082
0.1995	1.6557	101	0.4728	66.4516	19.7416
0.2291	1.6721	102	0.4730	66.8203	19.7949
0.2392	1.6885	103	0.4727	66.7281	19.8082
0.2617	1.7049	104	0.4724	66.7281	19.8748
0.3	1.7213	105	0.4725	66.6359	19.9014
0.2185	1.7377	106	0.4732	66.7281	20.1279
0.2271	1.7541	107	0.4735	66.6359	19.9147
0.1992	1.7705	108	0.4733	66.6359	19.9414
0.2123	1.7869	109	0.4738	66.6359	19.8615
0.2326	1.8033	110	0.4743	66.9124	19.8215
0.269	1.8197	111	0.4739	66.5438	19.7949
0.2053	1.8361	112	0.4739	66.7281	19.7815
0.2034	1.8525	113	0.4739	66.4516	19.7416
0.2033	1.8689	114	0.4739	66.4516	19.7682
0.2734	1.8852	115	0.4737	66.3594	19.7149
0.24	1.9016	116	0.4739	66.3594	19.7149
0.2612	1.9180	117	0.4737	66.6359	19.8082
0.2659	1.9344	118	0.4737	66.6359	19.8481
0.224	1.9508	119	0.4736	66.7281	19.8881
0.2204	1.9672	120	0.4736	66.8203	19.8215
0.2356	1.9836	121	0.4737	66.6359	19.7682
0.2193	2.0	122	0.4736	66.5438	19.8215

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_1.4e-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_1.4e-5