This model is a fine-tuned version of facebook/w2v-bert-2.0 on the CLEAR-GLOBAL/HAUSA_250_250H_YOURTTS - NA dataset. It achieves the following results on the evaluation set:

Loss: 0.2388
Wer: 0.3314
Cer: 0.1898

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

learning_rate: 3e-05
train_batch_size: 160
eval_batch_size: 160
seed: 42
distributed_type: multi-GPU
num_devices: 2
total_train_batch_size: 320
total_eval_batch_size: 320
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: linear
lr_scheduler_warmup_ratio: 0.1
num_epochs: 50.0
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer	Cer
0.5843	0.7225	1000	0.4228	0.4167	0.2123
0.2847	1.4451	2000	0.3274	0.3731	0.2006
0.0642	2.1676	3000	0.2891	0.3553	0.1963
0.082	2.8902	4000	0.2788	0.3512	0.1944
0.2373	3.6127	5000	0.2654	0.3462	0.1927
0.186	4.3353	6000	0.2633	0.3464	0.1930
0.0351	5.0578	7000	0.2603	0.3416	0.1921
0.0405	5.7803	8000	0.2670	0.3467	0.1932
0.1905	6.5029	9000	0.2480	0.3350	0.1904
0.1314	7.2254	10000	0.2616	0.3349	0.1907
0.1442	7.9480	11000	0.2596	0.3376	0.1914
0.0301	8.6705	12000	0.2439	0.3336	0.1897
0.1593	9.3931	13000	0.2410	0.3342	0.1898
0.1037	10.1156	14000	0.2461	0.3331	0.1905
0.0634	10.8382	15000	0.2598	0.3303	0.1896
0.0555	11.5607	16000	0.2400	0.3328	0.1899
0.1657	12.2832	17000	0.2437	0.3318	0.1896
0.0588	13.0058	18000	0.2386	0.3334	0.1904
0.058	13.7283	19000	0.2453	0.3294	0.1888
0.1711	14.4509	20000	0.2383	0.3317	0.1898
0.1441	15.1734	21000	0.2432	0.3324	0.1900
0.1502	15.8960	22000	0.2403	0.3287	0.1891
0.0333	16.6185	23000	0.2399	0.3313	0.1895
0.1604	17.3410	24000	0.2430	0.3273	0.1887
0.1162	18.0636	25000	0.2410	0.3277	0.1881

Framework versions

Transformers 4.48.1
Pytorch 2.7.0+cu126
Datasets 3.6.0
Tokenizers 0.21.1

CLEAR-Global/w2v-bert-2.0-hausa_250_250h_yourtts

w2v-bert-2.0-hausa_250_250h_yourtts-v2

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

CLEAR-Global/
w2v-bert-2.0-hausa_250_250h_yourtts