wav2vec2-large-xls-r-1B-swahili-cv-14

This model is a fine-tuned version of dmusingu/wav2vec2-large-xls-r-1B-swahili-cv-14 on the common_voice_14_0 dataset. It achieves the following results on the evaluation set:

• Loss: inf
• Wer: 0.2920
• Cer: 0.0924

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
• training_steps: 2000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer	Cer
0.3125	0.26	400	inf	0.3855	0.1231
0.3675	0.51	800	inf	0.4206	0.1332
0.3265	0.77	1200	inf	0.3514	0.1125
0.2924	1.02	1600	inf	0.3133	0.0998
0.2304	1.28	2000	inf	0.2920	0.0924

Framework versions

• Transformers 4.38.1
• Pytorch 2.2.1
• Datasets 2.17.0
• Tokenizers 0.15.2