This model is a fine-tuned version of Ycharo/w2v-bert-2.0-mongolian-colab-CV16.0 on the common_voice_16_0 dataset. It achieves the following results on the evaluation set:

Loss: 0.5129
Wer: 0.3230

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: 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: linear
lr_scheduler_warmup_steps: 500
num_epochs: 10
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
No log	1.1818	150	1.0154	0.7452
1.8517	2.3636	300	0.6427	0.5099
1.8517	3.5455	450	0.6015	0.4968
0.3476	4.7273	600	0.5711	0.4380
0.3476	5.9091	750	0.5515	0.3942
0.175	7.0870	900	0.5424	0.3599
0.175	8.2688	1050	0.5270	0.3390
0.0743	9.4506	1200	0.5129	0.3230

Framework versions

Transformers 4.50.3
Pytorch 2.6.0+cu124
Datasets 3.5.0
Tokenizers 0.21.1