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
1.1452	0.1087	5	1.0165	104.8699
0.9182	0.2174	10	0.8095	131.7380
0.7889	0.3261	15	0.6837	95.1301
0.6141	0.4348	20	0.6444	90.6801
0.6373	0.5435	25	0.6078	84.4668
0.611	0.6522	30	0.6075	83.1234
0.6567	0.7609	35	0.5800	86.5659
0.5233	0.8696	40	0.5562	75.2309
0.5035	0.9783	45	0.5548	79.5130
0.3647	1.0870	50	0.5488	80.1847
0.3553	1.1957	55	0.5533	78.8413
0.4018	1.3043	60	0.5588	77.3300
0.3974	1.4130	65	0.5613	82.7876
0.3017	1.5217	70	0.5459	74.9790
0.3	1.6304	75	0.5611	82.7876
0.4171	1.7391	80	0.5369	74.9790
0.3299	1.8478	85	0.5404	81.1923
0.3231	1.9565	90	0.5372	78.7573
0.2542	2.0652	95	0.5411	89.8405
0.2288	2.1739	100	0.5549	84.0470
0.2036	2.2826	105	0.5696	80.5206
0.2079	2.3913	110	0.5713	84.4668
0.1963	2.5	115	0.5760	81.7800
0.1938	2.6087	120	0.5701	78.5894
0.1771	2.7174	125	0.5723	81.7800
0.2393	2.8261	130	0.5706	86.3140
0.2395	2.9348	135	0.5641	75.6507
0.13	3.0435	140	0.5720	79.9328
0.112	3.1522	145	0.5805	84.4668
0.0956	3.2609	150	0.5961	83.3753
0.124	3.3696	155	0.6076	78.0856
0.1257	3.4783	160	0.6096	79.3451
0.1204	3.5870	165	0.6077	79.3451
0.1418	3.6957	170	0.6110	80.6045
0.1441	3.8043	175	0.6102	81.6961
0.1434	3.9130	180	0.6072	80.4366

nkkbr/whisper-large-v3-zatoichi-ja-EX-3-TRAIN_2_TO_10_EVAL_1

Whisper Large v3 - Japanese Zatoichi ASR