data2vec-large-uk

This model is a fine-tuned version of facebook/data2vec-audio-large-960h on the common_voice dataset. It achieves the following results on the evaluation set:

• eval_loss: 0.3472
• eval_wer: 0.3410
• eval_cer: 0.0832
• eval_runtime: 231.0008
• eval_samples_per_second: 25.108
• eval_steps_per_second: 3.139
• epoch: 33.06
• step: 20400

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: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 6
• total_train_batch_size: 48
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• num_epochs: 100
• mixed_precision_training: Native AMP

Framework versions

• Transformers 4.18.0
• Pytorch 1.11.0
• Datasets 1.18.3
• Tokenizers 0.12.1