You can deploy this model and start earning money today!

davidilag/
wav2vec2-xls-r-1b-distant-from-faroese-30-epochs_20250120_v6

Name: wav2vec2-xls-r-1b-distant-from-faroese-30-epochs_20250120_v6
Availability: InStock

Views

No views yet

wav2vec2-xls-r-1b-distant-from-faroese-30-epochs_20250120_v6

This model is a fine-tuned version of davidilag/wav2vec2-xls-r-1b-distant-from-faroese-30-epochs_20250120_v5 on an unknown dataset. It achieves the following results on the evaluation set:

More information needed

More information needed

More information needed

Lan=Turkish

The following hyperparameters were used during training:

learning_rate: 1e-05
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: 3000
num_epochs: 30
mixed_precision_training: Native AMP

Training Loss	Epoch	Step	Validation Loss	Wer	Cer
0.7602	1.2563	1000	0.3698	42.5303	9.6251
0.1885	2.5126	2000	0.1279	24.8181	5.3337
0.1845	3.7688	3000	0.1159	24.0238	5.0935
0.1442	5.0251	4000	0.1172	23.8906	5.3159
0.1279	6.2814	5000	0.1045	23.8193	5.1158
0.1032	7.5377	6000	0.1091	23.8573	4.9888
0.1787	8.7940	7000	0.1110	23.7004	4.8618
0.4532	10.0503	8000	0.3365	25.4174	5.0943
0.6745	11.3065	9000	0.4452	30.4590	5.9856
0.6692	12.5628	10000	0.4954	31.0012	6.0795
0.6029	13.8191	11000	0.4513	29.2128	5.7355
0.618	15.0754	12000	0.4681	29.5458	5.5915
0.6516	16.3317	13000	0.4594	32.2188	5.9064
0.6076	17.5879	14000	0.4576	32.4043	5.9325
0.6259	18.8442	15000	0.4569	32.4899	5.9487
0.6109	20.1005	16000	0.4569	32.4851	5.9487
0.5937	21.3568	17000	0.4569	32.4899	5.9471
0.5854	22.6131	18000	0.4569	32.4994	5.9495
0.5996	23.8693	19000	0.4569	32.4994	5.9495
0.6225	25.1256	20000	0.4569	32.4946	5.9487
0.6248	26.3819	21000	0.4569	32.4756	5.9464
0.6001	27.6382	22000	0.4569	32.4851	5.9471
0.6369	28.8945	23000	0.4568	32.4994	5.9502