wav2vec2-xlsr-53-espeak-cv-ft-sah-ntsema-colab

This model is a fine-tuned version of facebook/wav2vec2-xlsr-53-espeak-cv-ft on the audiofolder dataset. It achieves the following results on the evaluation set:

• Loss: 0.2143
• Wer: 0.2247

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: 4
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 8
• 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
2.7431	5.71	400	0.2879	0.4054
0.1876	11.42	800	0.2349	0.3023
0.0986	17.14	1200	0.2248	0.2701
0.0737	22.85	1600	0.2242	0.2428
0.0546	28.57	2000	0.2143	0.2247

Framework versions

• Transformers 4.24.0
• Pytorch 1.14.0.dev20221105+cu116
• Datasets 2.6.1
• Tokenizers 0.13.1