learning_rate: 0.0001
train_batch_size: 16
eval_batch_size: 8
seed: 42
gradient_accumulation_steps: 2
total_train_batch_size: 32
optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
lr_scheduler_type: cosine
lr_scheduler_warmup_steps: 6000
num_epochs: 30
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer	Cer
1.0778	0.6307	1000	0.6403	50.3070	15.8896
0.5579	1.2611	2000	0.3271	33.1503	9.3525
0.4943	1.8918	3000	0.2798	29.3342	8.2452
0.364	2.5222	4000	0.2706	29.1800	8.1400
0.3936	3.1526	5000	0.2704	29.8631	8.2974
0.3917	3.7833	6000	0.2831	30.0634	8.6321
0.401	4.4137	7000	0.2827	30.6442	8.7321
0.3556	5.0442	8000	0.2810	29.5504	8.3789
0.3758	5.6749	9000	0.2649	29.0661	8.2426
0.2998	6.3053	10000	0.2632	29.1123	8.1817
0.309	6.9360	11000	0.2585	28.3773	7.9461
0.327	7.5664	12000	0.2638	29.0359	8.1756
0.3288	8.1968	13000	0.2584	28.9768	8.0346
0.3118	8.8275	14000	0.2536	28.0905	7.8639
0.2333	9.4579	15000	0.2543	27.0745	7.5489
0.2406	10.0883	16000	0.2584	27.0399	7.6065
0.2323	10.7190	17000	0.2525	26.9203	7.4725
0.2726	11.3494	18000	0.2438	26.8165	7.4168
0.2361	11.9801	19000	0.2442	26.4202	7.3170
0.2217	12.6105	20000	0.2416	25.8078	7.0961
0.2006	13.2409	21000	0.2481	25.5728	7.0706
0.1744	13.8716	22000	0.2375	25.5901	6.9779
0.1803	14.5020	23000	0.2407	25.4806	6.9786
0.1835	15.1325	24000	0.2390	25.3941	6.9559
0.1835	15.7632	25000	0.2398	24.9474	6.8212
0.1356	16.3936	26000	0.2371	24.1332	6.5554
0.1275	17.0240	27000	0.2414	24.2600	6.5575
0.1354	17.6547	28000	0.2421	23.9026	6.4959
0.1377	18.2851	29000	0.2541	23.6561	6.3947
0.1245	18.9158	30000	0.2326	23.4256	6.3225
0.1146	19.5462	31000	0.2477	22.9918	6.1989
0.0842	20.1766	32000	0.2480	22.9082	6.1661
0.0822	20.8073	33000	0.2508	22.6992	6.0991
0.0813	21.4377	34000	0.2546	22.6776	6.0490
0.0876	22.0681	35000	0.2610	22.4600	5.9675
0.0906	22.6988	36000	0.2623	22.3274	5.9101
0.0755	23.3292	37000	0.2632	22.3750	5.8953
0.0734	23.9599	38000	0.2637	22.0378	5.8312
0.0665	24.5904	39000	0.2690	21.9455	5.7953
0.0628	25.2208	40000	0.2816	21.8274	5.7466
0.068	25.8515	41000	0.2767	21.8245	5.7178
0.0637	26.4819	42000	0.2801	21.7755	5.7103
0.0535	27.1123	43000	0.2787	21.7423	5.6853
0.0523	27.7430	44000	0.2833	21.6746	5.6771
0.0493	28.3734	45000	0.2822	21.6587	5.6658
0.0576	29.0038	46000	0.2805	21.6414	5.6632
0.0559	29.6345	47000	0.2819	21.6559	5.6618

Framework versions

Transformers 4.48.1
Pytorch 2.5.1+cu124
Datasets 3.2.0
Tokenizers 0.21.0

davidilag/wav2vec2-xls-r-1b-scandinavian-E3-100h-30-epochs-20250123

wav2vec2-xls-r-1b-scandinavian-E3-100h-30-epochs-20250123

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

davidilag/
wav2vec2-xls-r-1b-scandinavian-E3-100h-30-epochs-20250123