nkkbr/whisper-large-v3-zatoichi-ja-zatoichi-TEST-2-EX-3-TRAIN_2_TO_26_EVAL_1_COSINE

Whisper Large v3 - Japanese Zatoichi ASR

This model is a fine-tuned version of openai/whisper-large-v3 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.4476
• Wer: 71.9346

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: 1e-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
1.5378	0.0083	1	1.2957	98.1381
1.4921	0.0165	2	1.0248	96.0490
1.3743	0.0248	3	1.0124	95.3678
1.1751	0.0331	4	0.9946	93.8238
1.0236	0.0413	5	0.9759	90.6903
1.1897	0.0496	6	0.9466	87.9655
1.0615	0.0579	7	0.9063	87.0572
1.001	0.0661	8	0.8733	93.5513
0.9501	0.0744	9	0.8512	99.1826
1.0791	0.0826	10	0.8161	110.0363
0.9118	0.0909	11	0.7859	105.9946
0.8815	0.0992	12	0.7585	104.8592
0.8142	0.1074	13	0.7368	99.9546
0.8111	0.1157	14	0.7200	94.8229
0.7073	0.1240	15	0.6956	95.9582
0.8082	0.1322	16	0.6659	101.8165
0.7939	0.1405	17	0.6530	92.8701
0.7081	0.1488	18	0.6465	90.5540
0.6851	0.1570	19	0.6384	83.9691
0.7398	0.1653	20	0.6268	86.0127
0.7392	0.1736	21	0.6157	79.6094
0.6861	0.1818	22	0.6096	83.6512
0.6601	0.1901	23	0.6063	84.2416
0.634	0.1983	24	0.6011	85.8311
0.5911	0.2066	25	0.5976	89.4187
0.6306	0.2149	26	0.5950	91.7802
0.6915	0.2231	27	0.5912	93.3697
0.6969	0.2314	28	0.5853	88.4196
0.5674	0.2397	29	0.5856	85.0136
0.698	0.2479	30	0.5890	86.8756
0.8243	0.2562	31	0.5816	84.1508
0.5331	0.2645	32	0.5663	80.1544
0.6574	0.2727	33	0.5657	84.5595
0.6689	0.2810	34	0.5679	82.6521
0.6994	0.2893	35	0.5650	81.6076
0.6296	0.2975	36	0.5625	82.1072
0.6364	0.3058	37	0.5648	83.4696
0.5224	0.3140	38	0.5643	82.3342
0.5617	0.3223	39	0.5573	80.5631
0.6041	0.3306	40	0.5511	80.1090
0.5838	0.3388	41	0.5464	81.2443
0.6047	0.3471	42	0.5468	87.7384
0.4054	0.3554	43	0.5493	87.6022
0.4887	0.3636	44	0.5484	87.4659
0.5154	0.3719	45	0.5449	84.6957
0.405	0.3802	46	0.5399	80.1090
0.5478	0.3884	47	0.5367	79.0645
0.5762	0.3967	48	0.5341	78.2925
0.6441	0.4050	49	0.5309	75.7039
0.5237	0.4132	50	0.5272	72.3887
0.5383	0.4215	51	0.5240	77.7929
0.5002	0.4298	52	0.5240	77.4296
0.574	0.4380	53	0.5261	83.5604
0.5295	0.4463	54	0.5288	83.5150
0.5053	0.4545	55	0.5243	83.0154
0.5619	0.4628	56	0.5194	80.4269
0.6424	0.4711	57	0.5173	79.9728
0.5516	0.4793	58	0.5172	76.2035
0.5088	0.4876	59	0.5164	75.0681
0.5912	0.4959	60	0.5144	73.6149
0.5418	0.5041	61	0.5094	73.4787
0.5544	0.5124	62	0.5066	73.7965
0.5332	0.5207	63	0.5062	75.7493
0.5251	0.5289	64	0.5081	78.3379
0.4837	0.5372	65	0.5073	78.4287
0.458	0.5455	66	0.5047	78.2925
0.5927	0.5537	67	0.5043	79.6549
0.5764	0.5620	68	0.5055	75.5223
0.5646	0.5702	69	0.5039	75.6585
0.4158	0.5785	70	0.5005	74.7956
0.4543	0.5868	71	0.4959	76.4305
0.6407	0.5950	72	0.4928	74.6140
0.56	0.6033	73	0.4925	72.6158
0.5735	0.6116	74	0.4932	72.8883
0.5469	0.6198	75	0.4958	76.9755
0.4891	0.6281	76	0.4957	77.7929
0.5603	0.6364	77	0.4977	81.1535
0.5247	0.6446	78	0.5001	80.9718
0.595	0.6529	79	0.4982	79.9728
0.5471	0.6612	80	0.4926	78.4741
0.4362	0.6694	81	0.4867	76.8392
0.4606	0.6777	82	0.4840	78.1108
0.4938	0.6860	83	0.4831	82.2434
0.5463	0.6942	84	0.4834	80.9718
0.5244	0.7025	85	0.4819	79.7457
0.5682	0.7107	86	0.4807	81.9709
0.4386	0.7190	87	0.4792	80.0636
0.5093	0.7273	88	0.4798	82.9246
0.4358	0.7355	89	0.4805	85.6494
0.6347	0.7438	90	0.4780	77.9746
0.5476	0.7521	91	0.4744	79.3824
0.4541	0.7603	92	0.4718	83.6512
0.4316	0.7686	93	0.4706	85.3769
0.4403	0.7769	94	0.4693	84.2416
0.4217	0.7851	95	0.4681	84.1962
0.3985	0.7934	96	0.4674	84.0145
0.425	0.8017	97	0.4673	85.6494
0.5047	0.8099	98	0.4677	82.7884
0.4767	0.8182	99	0.4680	82.6521
0.495	0.8264	100	0.4698	83.4242
0.4793	0.8347	101	0.4708	82.9246
0.4736	0.8430	102	0.4704	80.9264
0.5275	0.8512	103	0.4700	79.9728
0.5839	0.8595	104	0.4704	78.8374
0.5803	0.8678	105	0.4704	76.8392
0.4352	0.8760	106	0.4688	75.5223
0.4732	0.8843	107	0.4672	73.8420
0.4922	0.8926	108	0.4661	72.3887
0.4906	0.9008	109	0.4657	72.7520
0.4649	0.9091	110	0.4659	69.0736
0.3851	0.9174	111	0.4666	71.0263
0.5232	0.9256	112	0.4667	71.5259
0.4163	0.9339	113	0.4662	71.9346
0.4634	0.9421	114	0.4666	72.6158
0.4259	0.9504	115	0.4648	72.6612
0.432	0.9587	116	0.4624	72.3887
0.5195	0.9669	117	0.4600	72.5250
0.5797	0.9752	118	0.4584	72.4796
0.4286	0.9835	119	0.4581	72.9791
0.4707	0.9917	120	0.4573	71.7530
0.4827	1.0	121	0.4563	71.9346
0.3233	1.0083	122	0.4554	71.1172
0.3977	1.0165	123	0.4539	70.3451
0.4133	1.0248	124	0.4524	69.6639
0.3025	1.0331	125	0.4516	69.2552
0.2876	1.0413	126	0.4527	70.4360
0.3792	1.0496	127	0.4535	70.5268
0.3335	1.0579	128	0.4560	71.7075
0.3322	1.0661	129	0.4584	72.8883
0.376	1.0744	130	0.4605	73.2516
0.3361	1.0826	131	0.4609	72.5704
0.3784	1.0909	132	0.4607	72.8429
0.3134	1.0992	133	0.4605	71.2988
0.3392	1.1074	134	0.4620	70.7084
0.269	1.1157	135	0.4636	70.0727
0.3121	1.1240	136	0.4656	69.7094
0.268	1.1322	137	0.4667	70.2089
0.3238	1.1405	138	0.4661	71.2080
0.3296	1.1488	139	0.4656	71.9800
0.3074	1.1570	140	0.4670	73.6149
0.3581	1.1653	141	0.4677	74.0236
0.324	1.1736	142	0.4664	73.9782
0.3254	1.1818	143	0.4650	73.9782
0.3061	1.1901	144	0.4630	73.9328
0.2935	1.1983	145	0.4602	73.1153
0.3057	1.2066	146	0.4574	72.4342
0.2668	1.2149	147	0.4569	71.0718
0.2705	1.2231	148	0.4561	70.7539
0.3467	1.2314	149	0.4563	70.6630
0.3091	1.2397	150	0.4565	71.3442
0.2788	1.2479	151	0.4570	71.3896
0.3339	1.2562	152	0.4570	71.5713
0.3077	1.2645	153	0.4575	71.7075
0.3156	1.2727	154	0.4576	71.4805
0.3102	1.2810	155	0.4572	71.5259
0.2836	1.2893	156	0.4568	73.2516
0.2536	1.2975	157	0.4572	72.0254
0.3354	1.3058	158	0.4582	72.4796
0.3433	1.3140	159	0.4590	72.9791
0.3636	1.3223	160	0.4601	73.0245
0.2483	1.3306	161	0.4601	73.1608
0.2769	1.3388	162	0.4597	73.2970
0.3033	1.3471	163	0.4591	73.4332
0.2952	1.3554	164	0.4582	73.1153
0.2569	1.3636	165	0.4568	72.8883
0.2845	1.3719	166	0.4552	72.3887
0.3287	1.3802	167	0.4537	71.6621
0.4109	1.3884	168	0.4533	71.2534
0.3929	1.3967	169	0.4526	70.8447
0.3329	1.4050	170	0.4522	70.7084
0.2335	1.4132	171	0.4519	70.7539
0.2875	1.4215	172	0.4516	72.3433
0.3015	1.4298	173	0.4517	72.6158
0.3022	1.4380	174	0.4521	73.3424
0.3073	1.4463	175	0.4529	72.8429
0.3425	1.4545	176	0.4534	72.8883
0.3096	1.4628	177	0.4540	73.0245
0.2332	1.4711	178	0.4547	73.1608
0.3496	1.4793	179	0.4548	73.5241
0.3898	1.4876	180	0.4547	72.8429
0.2823	1.4959	181	0.4537	72.5704
0.3001	1.5041	182	0.4533	72.4342
0.3317	1.5124	183	0.4527	72.6158
0.3004	1.5207	184	0.4528	72.7066
0.3719	1.5289	185	0.4522	72.1163
0.3361	1.5372	186	0.4518	72.4796
0.2271	1.5455	187	0.4522	73.0699
0.2849	1.5537	188	0.4524	73.7057
0.387	1.5620	189	0.4525	73.8420
0.3307	1.5702	190	0.4521	74.1599
0.3349	1.5785	191	0.4520	73.5695
0.2609	1.5868	192	0.4517	73.6603
0.2763	1.5950	193	0.4516	73.0245
0.2667	1.6033	194	0.4509	72.7520
0.2721	1.6116	195	0.4505	72.6612
0.2611	1.6198	196	0.4502	72.5704
0.262	1.6281	197	0.4497	73.3424
0.2399	1.6364	198	0.4495	73.3878
0.2943	1.6446	199	0.4491	73.2516
0.3182	1.6529	200	0.4489	73.0245
0.2951	1.6612	201	0.4485	72.6612
0.2931	1.6694	202	0.4482	72.5704
0.2867	1.6777	203	0.4478	72.4342
0.3684	1.6860	204	0.4478	71.7075
0.3363	1.6942	205	0.4473	71.7984
0.3265	1.7025	206	0.4477	72.3433
0.2725	1.7107	207	0.4476	72.4342
0.3738	1.7190	208	0.4476	72.3433
0.3246	1.7273	209	0.4476	72.0708
0.2749	1.7355	210	0.4477	72.6158
0.282	1.7438	211	0.4479	72.7520
0.3303	1.7521	212	0.4482	72.9791
0.3164	1.7603	213	0.4483	72.6158
0.2679	1.7686	214	0.4484	72.8429
0.2838	1.7769	215	0.4481	72.8429
0.4008	1.7851	216	0.4482	72.7066
0.3147	1.7934	217	0.4483	72.5704
0.3461	1.8017	218	0.4483	72.4796
0.3171	1.8099	219	0.4482	72.3433
0.3299	1.8182	220	0.4481	72.2525
0.2426	1.8264	221	0.4480	72.2979
0.2564	1.8347	222	0.4481	72.1617
0.283	1.8430	223	0.4478	72.0708
0.29	1.8512	224	0.4479	71.8438
0.3392	1.8595	225	0.4478	72.2071
0.2896	1.8678	226	0.4478	72.3433
0.2383	1.8760	227	0.4481	72.2979
0.2621	1.8843	228	0.4478	72.0254
0.3141	1.8926	229	0.4474	72.2525
0.2319	1.9008	230	0.4476	72.0708
0.2858	1.9091	231	0.4477	72.1617
0.3397	1.9174	232	0.4477	72.1163
0.4042	1.9256	233	0.4476	71.9346
0.3133	1.9339	234	0.4477	72.3433
0.3107	1.9421	235	0.4477	71.9346
0.2506	1.9504	236	0.4477	71.9800
0.2688	1.9587	237	0.4476	71.9346
0.3431	1.9669	238	0.4474	72.2071
0.27	1.9752	239	0.4475	72.0708
0.3266	1.9835	240	0.4480	72.1617
0.3253	1.9917	241	0.4477	72.0708
0.2657	2.0	242	0.4476	71.9346

Framework versions

• Transformers 4.57.3
• Pytorch 2.9.1+cu128
• Datasets 4.4.1
• Tokenizers 0.22.1