wav2vec2-large-xls-r-1b-asante-twi

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

• Loss: 0.4512
• Wer: 0.3632

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: 40
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.0587	5.8394	400	0.4475	0.4494
0.1148	11.6788	800	0.4678	0.4175
0.0574	17.5182	1200	0.4135	0.4124
0.0422	23.3577	1600	0.4535	0.3791
0.0269	29.1971	2000	0.4690	0.3693
0.0198	35.0365	2400	0.4512	0.3632

Framework versions

• Transformers 4.45.2
• Pytorch 2.5.1+cu121
• Datasets 3.1.0
• Tokenizers 0.20.3