vtt-indonesia

This model is a fine-tuned version of facebook/wav2vec2-xls-r-300m on the common_voice dataset. It achieves the following results on the evaluation set:

• Loss: 0.3472
• Wer: 0.3582

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: 0.0003
• train_batch_size: 16
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 32
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• num_epochs: 30
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
3.7612	3.23	400	0.6405	0.6714
0.4143	6.45	800	0.3772	0.4974
0.2068	9.68	1200	0.3877	0.4442
0.1436	12.9	1600	0.3785	0.4212
0.1133	16.13	2000	0.3944	0.4144
0.09	19.35	2400	0.3695	0.3925
0.0705	22.58	2800	0.3706	0.3846
0.057	25.81	3200	0.3720	0.3725
0.048	29.03	3600	0.3472	0.3582

Framework versions

• Transformers 4.11.3
• Pytorch 1.10.0+cu113
• Datasets 1.18.3
• Tokenizers 0.10.3