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
1.3324	0.0161	1	1.2957	98.1381
1.4072	0.0323	2	1.0274	96.1853
1.2085	0.0484	3	1.0033	95.2316
1.1135	0.0645	4	0.9624	92.1889
1.0292	0.0806	5	0.9335	88.1926
0.8442	0.0968	6	0.8707	91.7802
0.9434	0.1129	7	0.8318	90.9628
0.771	0.1290	8	0.7904	103.3606
0.8658	0.1452	9	0.7604	104.2688
0.6908	0.1613	10	0.7239	106.4487
0.8142	0.1774	11	0.6971	99.2734
0.7678	0.1935	12	0.6638	96.0036
0.8074	0.2097	13	0.6442	94.2325
0.7395	0.2258	14	0.6306	90.5540
0.6385	0.2419	15	0.6205	86.5577
0.5888	0.2581	16	0.6113	85.3315
0.488	0.2742	17	0.6073	87.0572
0.6574	0.2903	18	0.5998	81.1081
0.4911	0.3065	19	0.5923	85.5132
0.5855	0.3226	20	0.5920	91.0082
0.5868	0.3387	21	0.5827	86.6939
0.6491	0.3548	22	0.5780	87.1026
0.5895	0.3710	23	0.5769	88.1926
0.5515	0.3871	24	0.5705	88.3742
0.4251	0.4032	25	0.5650	81.2443
0.5516	0.4194	26	0.5701	83.9237
0.5925	0.4355	27	0.5747	83.2879
0.5876	0.4516	28	0.5703	81.5168
0.5373	0.4677	29	0.5541	79.3370
0.3455	0.4839	30	0.5490	77.3388
0.4579	0.5	31	0.5481	75.2498
0.4227	0.5161	32	0.5450	75.2498
0.4522	0.5323	33	0.5398	75.4314
0.4462	0.5484	34	0.5406	78.2016
0.5487	0.5645	35	0.5438	78.4287
0.4535	0.5806	36	0.5492	82.1526
0.5018	0.5968	37	0.5509	84.2870
0.5043	0.6129	38	0.5467	83.3333
0.457	0.6290	39	0.5417	82.1072
0.3447	0.6452	40	0.5368	82.5159
0.3967	0.6613	41	0.5315	78.7012
0.4523	0.6774	42	0.5274	78.0654
0.3704	0.6935	43	0.5266	80.7448
0.4888	0.7097	44	0.5274	77.6113
0.4767	0.7258	45	0.5295	79.0191
0.4144	0.7419	46	0.5302	78.9282
0.4575	0.7581	47	0.5300	81.9709
0.472	0.7742	48	0.5294	82.7884
0.4272	0.7903	49	0.5291	80.6540
0.369	0.8065	50	0.5262	78.6104
0.342	0.8226	51	0.5255	80.5631
0.4054	0.8387	52	0.5265	78.1108
0.3719	0.8548	53	0.5259	77.3842
0.3941	0.8710	54	0.5258	78.3833
0.3618	0.8871	55	0.5277	80.5177
0.3531	0.9032	56	0.5275	81.3806
0.4127	0.9194	57	0.5232	78.5195
0.3053	0.9355	58	0.5180	76.3851
0.3006	0.9516	59	0.5148	74.7956
0.3529	0.9677	60	0.5129	74.4323
0.3644	0.9839	61	0.5125	77.4750
0.3502	1.0	62	0.5127	77.9292
0.2419	1.0161	63	0.5142	76.6576
0.2884	1.0323	64	0.5181	79.3824
0.2684	1.0484	65	0.5216	82.0163
0.1972	1.0645	66	0.5246	82.3797
0.2046	1.0806	67	0.5253	81.8801
0.2047	1.0968	68	0.5265	80.6540
0.212	1.1129	69	0.5301	80.3361
0.298	1.1290	70	0.5324	80.2452
0.2458	1.1452	71	0.5358	80.8356
0.194	1.1613	72	0.5391	79.7457
0.2058	1.1774	73	0.5415	78.9737
0.3015	1.1935	74	0.5415	79.0191
0.2601	1.2097	75	0.5401	77.9292
0.223	1.2258	76	0.5372	76.4759
0.246	1.2419	77	0.5353	75.1589
0.2457	1.2581	78	0.5344	74.8865
0.2259	1.2742	79	0.5335	75.9764
0.2098	1.2903	80	0.5322	76.2943
0.2211	1.3065	81	0.5312	76.8392
0.2625	1.3226	82	0.5319	81.1535
0.2211	1.3387	83	0.5326	82.8792
0.2411	1.3548	84	0.5341	81.6530
0.2954	1.3710	85	0.5348	81.6985
0.183	1.3871	86	0.5375	81.0627
0.2034	1.4032	87	0.5387	81.5622
0.1844	1.4194	88	0.5384	81.1535
0.2199	1.4355	89	0.5362	79.9273
0.1877	1.4516	90	0.5335	79.9728
0.2081	1.4677	91	0.5308	77.7021
0.2281	1.4839	92	0.5288	79.1099
0.1985	1.5	93	0.5270	78.6558
0.2101	1.5161	94	0.5264	78.3379
0.258	1.5323	95	0.5253	76.6122
0.217	1.5484	96	0.5250	76.8847
0.2234	1.5645	97	0.5251	77.4296
0.2638	1.5806	98	0.5252	77.9746
0.2384	1.5968	99	0.5253	80.0182
0.1946	1.6129	100	0.5249	80.1090
0.2107	1.6290	101	0.5250	78.4287
0.2251	1.6452	102	0.5249	80.3815
0.2894	1.6613	103	0.5246	80.0182
0.2127	1.6774	104	0.5248	79.7911
0.229	1.6935	105	0.5247	79.7457
0.2061	1.7097	106	0.5246	80.0636
0.1713	1.7258	107	0.5248	79.7003
0.2129	1.7419	108	0.5253	80.9264
0.2129	1.7581	109	0.5255	79.7003
0.1348	1.7742	110	0.5257	77.9292
0.2143	1.7903	111	0.5258	77.9746
0.2351	1.8065	112	0.5258	78.1108
0.2325	1.8226	113	0.5257	77.8837
0.1948	1.8387	114	0.5259	77.7475
0.1935	1.8548	115	0.5257	77.7021
0.1859	1.8710	116	0.5260	78.0200
0.1493	1.8871	117	0.5260	78.0200
0.2198	1.9032	118	0.5259	77.7929
0.1952	1.9194	119	0.5261	78.1108
0.1684	1.9355	120	0.5260	77.9746
0.176	1.9516	121	0.5258	77.9292
0.2109	1.9677	122	0.5259	78.1108
0.222	1.9839	123	0.5260	77.7475
0.1901	2.0	124	0.5262	77.9292

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-EX-12-TRAIN_2_TO_10_EVAL_1_COSINE

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-EX-12-TRAIN_2_TO_10_EVAL_1_COSINE