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 results

Training Loss	Epoch	Step	Validation Loss	Wer
1.2597	0.0575	5	1.1471	102.0476
1.0355	0.1149	10	0.9728	141.0364
0.9717	0.1724	15	0.8136	131.1743
0.6622	0.2299	20	0.7011	101.5462
0.7253	0.2874	25	0.6469	87.8813
0.6762	0.3448	30	0.5931	86.1262
0.5704	0.4023	35	0.5743	84.9979
0.5909	0.4598	40	0.5359	81.4877
0.514	0.5172	45	0.5195	83.3264
0.5108	0.5747	50	0.5025	83.3682
0.493	0.6322	55	0.4844	73.2135
0.3642	0.6897	60	0.4618	76.4313
0.4218	0.7471	65	0.4490	69.9540
0.4934	0.8046	70	0.4393	68.0735
0.327	0.8621	75	0.4175	63.8529
0.4492	0.9195	80	0.4042	74.7179
0.3388	0.9770	85	0.3918	69.3272
0.2804	1.0345	90	0.3848	74.5926
0.2821	1.0920	95	0.3784	64.2290
0.2451	1.1494	100	0.3742	60.7606
0.2684	1.2069	105	0.3705	58.7129
0.3079	1.2644	110	0.3624	57.9607
0.3265	1.3218	115	0.3616	55.6623
0.2793	1.3793	120	0.3568	59.2562
0.26	1.4368	125	0.3460	67.3631
0.2662	1.4943	130	0.3445	55.5370
0.2394	1.5517	135	0.3378	66.8617
0.2378	1.6092	140	0.3304	58.6711
0.2334	1.6667	145	0.3237	60.8859
0.1707	1.7241	150	0.3175	52.7372
0.1966	1.7816	155	0.3098	46.1346
0.285	1.8391	160	0.3040	51.9432
0.1576	1.8966	165	0.2996	44.4212
0.1829	1.9540	170	0.2931	47.5972
0.101	2.0115	175	0.2927	50.6477
0.1187	2.0690	180	0.2872	48.8090
0.0925	2.1264	185	0.2976	44.0033
0.1119	2.1839	190	0.2885	45.3406
0.1418	2.2414	195	0.2879	43.8780
0.1252	2.2989	200	0.2803	50.5224
0.1713	2.3563	205	0.2745	40.2424
0.1273	2.4138	210	0.2696	41.3707
0.1265	2.4713	215	0.2725	42.0811
0.1117	2.5287	220	0.2691	40.2842
0.0901	2.5862	225	0.2641	42.6661
0.1157	2.6437	230	0.2614	35.8964
0.0945	2.7011	235	0.2597	37.9022
0.0844	2.7586	240	0.2589	40.5349
0.1705	2.8161	245	0.2567	38.8634
0.1272	2.8736	250	0.2513	39.1141
0.0856	2.9310	255	0.2500	37.7351
0.1073	2.9885	260	0.2446	36.0217
0.0468	3.0460	265	0.2424	33.2219
0.0439	3.1034	270	0.2470	33.4726
0.0488	3.1609	275	0.2461	32.3025
0.0392	3.2184	280	0.2457	30.6728
0.0379	3.2759	285	0.2503	30.7982
0.0462	3.3333	290	0.2471	32.1772
0.0456	3.3908	295	0.2429	31.3414
0.0522	3.4483	300	0.2414	32.8876
0.0419	3.5057	305	0.2430	34.4338
0.0397	3.5632	310	0.2422	34.7681
0.046	3.6207	315	0.2404	32.9294
0.0668	3.6782	320	0.2403	33.2637
0.045	3.7356	325	0.2404	29.1266
0.0443	3.7931	330	0.2391	28.6252
0.0571	3.8506	335	0.2364	28.5416
0.0617	3.9080	340	0.2345	28.2908
0.0438	3.9655	345	0.2336	27.9983

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-8

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-8