English_ASR

This model is a fine-tuned version of facebook/wav2vec2-base on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.4971
• Wer: 0.3397

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: 0.0001
• train_batch_size: 32
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 1000
• num_epochs: 30
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
3.3432	4.0	500	1.1711	0.7767
0.5691	8.0	1000	0.4613	0.4357
0.2182	12.0	1500	0.4715	0.3853
0.1267	16.0	2000	0.4307	0.3607
0.0846	20.0	2500	0.4971	0.3537
0.0608	24.0	3000	0.4712	0.3419
0.0457	28.0	3500	0.4971	0.3397

Framework versions

• Transformers 4.11.3
• Pytorch 1.9.1
• Datasets 1.13.3
• Tokenizers 0.10.3