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: linear
lr_scheduler_warmup_ratio: 0.1
num_epochs: 4

Training Loss	Epoch	Step	Validation Loss	Wer
1.0434	0.1515	5	0.9573	87.9141
0.7689	0.3030	10	0.7493	99.8209
0.571	0.4545	15	0.6325	97.1352
0.6033	0.6061	20	0.6160	93.2856
0.6603	0.7576	25	0.5735	76.8129
0.6532	0.9091	30	0.5725	82.6321
0.4217	1.0606	35	0.5513	75.9176
0.3452	1.2121	40	0.5515	79.2301
0.3774	1.3636	45	0.5454	79.5882
0.3132	1.5152	50	0.5429	80.2149
0.3425	1.6667	55	0.5302	77.4396
0.2999	1.8182	60	0.5265	77.1710
0.3142	1.9697	65	0.5210	77.7977
0.1643	2.1212	70	0.5202	74.2167
0.2764	2.2727	75	0.5426	77.5291
0.2203	2.4242	80	0.5476	75.2014
0.2139	2.5758	85	0.5462	77.7977
0.2068	2.7273	90	0.5367	76.1862
0.2348	2.8788	95	0.5332	75.1119
0.1417	3.0303	100	0.5425	81.5577
0.1538	3.1818	105	0.5491	76.4548
0.1353	3.3333	110	0.5589	75.2014
0.1365	3.4848	115	0.5716	78.7825
0.1403	3.6364	120	0.5696	76.7234
0.1221	3.7879	125	0.5637	75.3805
0.1217	3.9394	130	0.5615	75.1119

nkkbr/whisper-large-v3-zatoichi-ja-EX-5-TRAIN_2_TO_10_EVAL_1

Whisper Large v3 - Japanese Zatoichi ASR