learning_rate: 1e-05
train_batch_size: 64
eval_batch_size: 64
seed: 42
gradient_accumulation_steps: 2
total_train_batch_size: 128
optimizer: Use adamw_torch 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: 0.06
num_epochs: 1
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Model Preparation Time	Wer Ortho	Wer
0.1188	0.0188	32	0.6057	0.0068	30.9517	13.9902
0.0922	0.0377	64	0.6032	0.0068	31.4139	14.9377
0.0972	0.0565	96	0.6382	0.0068	31.6199	15.0370
0.0993	0.0754	128	0.6267	0.0068	32.5093	15.9498
0.0892	0.0942	160	0.6853	0.0068	32.3234	15.4884
0.0909	0.1131	192	0.6374	0.0068	32.0822	15.3346
0.096	0.1319	224	0.6838	0.0068	32.1827	15.5876
0.1013	0.1508	256	0.6491	0.0068	31.4692	15.0568
0.0927	0.1696	288	0.6646	0.0068	32.0169	15.4785
0.0815	0.1885	320	0.6886	0.0068	31.7657	15.1362
0.0852	0.2073	352	0.6637	0.0068	31.4792	14.8832
0.0809	0.2261	384	0.6726	0.0068	31.7606	15.2255
0.0826	0.2450	416	0.6461	0.0068	31.9013	15.1263
0.0822	0.2638	448	0.6735	0.0068	31.3285	14.9080
0.0796	0.2827	480	0.6744	0.0068	31.3838	14.7343
0.0751	0.3015	512	0.6731	0.0068	31.5446	14.8286
0.0817	0.3204	544	0.6401	0.0068	31.5848	14.8435
0.0793	0.3392	576	0.6894	0.0068	31.9465	15.1362
0.0725	0.3581	608	0.6499	0.0068	31.5245	14.6202
0.0748	0.3769	640	0.6746	0.0068	31.3737	14.5012
0.0778	0.3958	672	0.6622	0.0068	31.5295	14.5111
0.0703	0.4146	704	0.6969	0.0068	31.7204	14.7591
0.074	0.4335	736	0.6794	0.0068	31.3486	14.5954
0.0705	0.4523	768	0.6916	0.0068	31.2682	14.2482
0.0756	0.4711	800	0.6822	0.0068	31.5044	14.6252
0.0824	0.4900	832	0.6823	0.0068	30.8713	14.0299
0.0707	0.5088	864	0.6938	0.0068	30.9567	14.0795
0.07	0.5277	896	0.6576	0.0068	30.9969	14.1192
0.0796	0.5465	928	0.6691	0.0068	31.0873	14.0051
0.0677	0.5654	960	0.6846	0.0068	30.7557	13.9703
0.0713	0.5842	992	0.6493	0.0068	30.4793	13.7769
0.0668	0.6031	1024	0.6660	0.0068	30.5597	13.8314
0.0636	0.6219	1056	0.6407	0.0068	30.5396	13.4593
0.0667	0.6408	1088	0.6685	0.0068	30.6854	13.6429
0.0622	0.6596	1120	0.6446	0.0068	30.2231	13.3998
0.062	0.6784	1152	0.6523	0.0068	30.3085	13.3651
0.0732	0.6973	1184	0.6514	0.0068	30.2231	13.1815
0.0697	0.7161	1216	0.6550	0.0068	30.1779	13.2758
0.0629	0.7350	1248	0.6495	0.0068	30.1126	13.0327
0.0655	0.7538	1280	0.6371	0.0068	30.1779	13.1766
0.0659	0.7727	1312	0.6468	0.0068	30.0623	12.9236
0.0672	0.7915	1344	0.6535	0.0068	29.8261	12.8987
0.0596	0.8104	1376	0.6580	0.0068	29.9719	12.8045
0.0718	0.8292	1408	0.6536	0.0068	29.8563	12.7946
0.0623	0.8481	1440	0.6609	0.0068	29.8312	12.6457
0.0617	0.8669	1472	0.6639	0.0068	29.5347	12.3580
0.0624	0.8857	1504	0.6570	0.0068	29.5448	12.3332
0.0627	0.9046	1536	0.6650	0.0068	29.7608	12.5267
0.063	0.9234	1568	0.6508	0.0068	29.6402	12.4820
0.0604	0.9423	1600	0.6574	0.0068	29.5096	12.3381
0.0655	0.9611	1632	0.6654	0.0068	29.5146	12.3332
0.0626	0.9800	1664	0.6571	0.0068	29.5799	12.4026
0.0609	0.9988	1696	0.6574	0.0068	29.5297	12.3977

Framework versions

Transformers 4.51.3
Pytorch 2.5.1
Datasets 3.6.0
Tokenizers 0.21.1

SamuelPfisterer1/whisper-large-v3-turbo-danish-denmark

whisper-large-v3-turbo-danish-denmark

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

SamuelPfisterer1/
whisper-large-v3-turbo-danish-denmark