wav2vec2-large-xls-r-300m-asante-twi

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

• Loss: 0.4944
• Wer: 0.4114

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
3.4512	5.8394	400	0.5347	0.6223
0.2568	11.6788	800	0.4314	0.4948
0.1053	17.5182	1200	0.4666	0.4564
0.0579	23.3577	1600	0.4916	0.4339
0.0395	29.1971	2000	0.4894	0.4203
0.0274	35.0365	2400	0.4944	0.4114

Framework versions

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