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
1.6926	0.0156	1	1.5142	97.9390
1.5649	0.0312	2	1.1615	94.3941
1.2273	0.0469	3	1.1346	92.8277
1.1525	0.0625	4	1.0895	98.5161
1.1459	0.0781	5	1.0545	87.8813
1.1769	0.0938	6	0.9801	102.0610
1.0835	0.1094	7	0.9292	123.5779
1.031	0.125	8	0.8885	136.9332
0.8603	0.1406	9	0.8583	118.3017
0.9318	0.1562	10	0.8179	106.7601
0.8396	0.1719	11	0.7828	107.3372
0.8065	0.1875	12	0.7552	85.1608
0.8381	0.2031	13	0.7265	91.5911
0.8135	0.2188	14	0.7102	102.6381
0.782	0.2344	15	0.6914	90.7667
0.8107	0.25	16	0.6814	81.9456
0.6933	0.2656	17	0.6737	81.0387
0.7648	0.2812	18	0.6693	80.1319
0.7436	0.2969	19	0.6642	81.5334
0.7237	0.3125	20	0.6576	82.7700
0.7152	0.3281	21	0.6533	97.1970
0.58	0.3438	22	0.6489	91.5911
0.6722	0.3594	23	0.6412	91.4262
0.6033	0.375	24	0.6357	88.1286
0.7329	0.3906	25	0.6313	87.5515
0.571	0.4062	26	0.6271	86.1500
0.703	0.4219	27	0.6347	77.6587
0.5161	0.4375	28	0.6375	86.5622
0.6749	0.4531	29	0.6309	94.7238
0.7403	0.4688	30	0.6214	98.6810
0.5672	0.4844	31	0.6159	93.4872
0.5892	0.5	32	0.6141	93.9819
0.5562	0.5156	33	0.6148	92.4155
0.6753	0.5312	34	0.6100	84.8310
0.6111	0.5469	35	0.6057	84.5012
0.7324	0.5625	36	0.6053	85.5730
0.5728	0.5781	37	0.6064	86.3149
0.6259	0.5938	38	0.6108	93.8170
0.6776	0.6094	39	0.6119	87.2218
0.4643	0.625	40	0.6087	91.6735
0.5212	0.6406	41	0.6056	90.7667
0.5659	0.6562	42	0.6008	90.6018
0.6181	0.6719	43	0.5955	94.4765
0.646	0.6875	44	0.5935	92.4979
0.6606	0.7031	45	0.5942	91.5911
0.6496	0.7188	46	0.5961	92.0033
0.5046	0.7344	47	0.5955	92.3331
0.6341	0.75	48	0.5949	92.5804
0.5994	0.7656	49	0.5968	96.3726
0.5887	0.7812	50	0.5994	90.3545
0.6751	0.7969	51	0.6019	92.2506
0.5823	0.8125	52	0.5998	92.6628
0.5769	0.8281	53	0.5940	94.6414
0.5419	0.8438	54	0.5856	92.4155
0.5484	0.8594	55	0.5769	88.4584
0.7178	0.875	56	0.5711	87.7164
0.5123	0.8906	57	0.5671	87.1393
0.595	0.9062	58	0.5647	83.8417
0.5813	0.9219	59	0.5646	83.7593
0.5636	0.9375	60	0.5645	80.3792
0.5669	0.9531	61	0.5636	80.9563
0.4546	0.9688	62	0.5654	81.2036
0.5659	0.9844	63	0.5651	82.9349
0.6663	1.0	64	0.5676	79.8021
0.4874	1.0156	65	0.5693	81.8631
0.3302	1.0312	66	0.5687	81.9456
0.3229	1.0469	67	0.5677	81.8631
0.4085	1.0625	68	0.5665	88.7881
0.3832	1.0781	69	0.5658	90.1072
0.3247	1.0938	70	0.5649	87.2218
0.3233	1.1094	71	0.5653	86.8096
0.3361	1.125	72	0.5656	85.8203
0.4531	1.1406	73	0.5664	88.7057
0.3892	1.1562	74	0.5673	85.5730
0.3162	1.1719	75	0.5665	85.6554
0.3259	1.1875	76	0.5662	85.3256
0.3881	1.2031	77	0.5666	88.2110
0.2689	1.2188	78	0.5672	88.6232
0.4077	1.2344	79	0.5668	94.1467
0.4236	1.25	80	0.5666	95.3009
0.3049	1.2656	81	0.5666	83.5944
0.3408	1.2812	82	0.5669	83.8417
0.3555	1.2969	83	0.5677	84.1715
0.3359	1.3125	84	0.5677	78.3182
0.3599	1.3281	85	0.5671	78.3182
0.34	1.3438	86	0.5653	77.1641
0.3018	1.3594	87	0.5628	76.5045
0.4373	1.375	88	0.5606	80.2143
0.3763	1.3906	89	0.5595	79.3899
0.2653	1.4062	90	0.5587	87.9637
0.3073	1.4219	91	0.5590	87.4691
0.2592	1.4375	92	0.5584	75.5977
0.451	1.4531	93	0.5581	79.3075
0.3357	1.4688	94	0.5579	87.9637
0.4141	1.4844	95	0.5590	88.3759
0.4035	1.5	96	0.5594	89.0354
0.3463	1.5156	97	0.5602	89.7774
0.2969	1.5312	98	0.5611	87.0569
0.3413	1.5469	99	0.5623	87.5515
0.3787	1.5625	100	0.5630	87.9637
0.336	1.5781	101	0.5640	88.7057
0.3305	1.5938	102	0.5648	89.6125
0.3624	1.6094	103	0.5649	82.1929
0.3873	1.625	104	0.5652	82.6051
0.4136	1.6406	105	0.5651	82.4402
0.3165	1.6562	106	0.5647	83.1822
0.3418	1.6719	107	0.5639	82.4402
0.3321	1.6875	108	0.5627	81.9456
0.3161	1.7031	109	0.5615	89.8599
0.325	1.7188	110	0.5605	89.6125
0.4297	1.7344	111	0.5594	89.6125
0.3513	1.75	112	0.5586	89.4477
0.2977	1.7656	113	0.5579	88.7057
0.3276	1.7812	114	0.5570	88.1286
0.4109	1.7969	115	0.5562	88.2935
0.335	1.8125	116	0.5561	87.4691
0.2993	1.8281	117	0.5558	87.2218
0.3826	1.8438	118	0.5551	86.8920
0.3004	1.8594	119	0.5552	86.8920
0.2935	1.875	120	0.5547	87.3042
0.3617	1.8906	121	0.5546	87.2218
0.276	1.9062	122	0.5546	86.8920
0.3136	1.9219	123	0.5544	86.9744
0.3427	1.9375	124	0.5547	86.8920
0.3652	1.9531	125	0.5546	86.7271
0.2848	1.9688	126	0.5542	86.8920
0.3193	1.9844	127	0.5543	87.0569
0.3492	2.0	128	0.5542	86.8096

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-2-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-2-TRAIN_2_TO_10_EVAL_1_COSINE