This model is a fine-tuned version of facebook/w2v-bert-2.0 on the common_voice_16_0 dataset. It achieves the following results on the evaluation set:

Loss: 0.2582
Wer: 0.1707

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: 5e-05
train_batch_size: 32
eval_batch_size: 8
seed: 42
gradient_accumulation_steps: 2
total_train_batch_size: 64
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_steps: 500
num_epochs: 20
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
1.8342	3.1915	300	0.2208	0.2664
0.1353	6.3830	600	0.2000	0.2162
0.0657	9.5745	900	0.1967	0.1965
0.0284	12.7660	1200	0.2242	0.1858
0.008	15.9574	1500	0.2424	0.1744
0.0021	19.1489	1800	0.2582	0.1707

Framework versions

Transformers 4.53.0.dev0
Pytorch 2.7.1+cu126
Datasets 2.15.0
Tokenizers 0.21.1

Sargis001/w2v-bert-2.0-armenian-CV16.0

w2v-bert-2.0-armenian-CV16.0

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

Sargis001/
w2v-bert-2.0-armenian-CV16.0