working

This model is a fine-tuned version of facebook/wav2vec2-base on the vivos dataset. It achieves the following results on the evaluation set:

• Loss: 0.7102
• Wer: 0.4126

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
• gradient_accumulation_steps: 2
• total_train_batch_size: 64
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• num_epochs: 20
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
6.3485	2.0	292	3.7183	1.0
3.4479	4.0	584	3.5977	1.0
2.948	6.0	876	1.7093	0.8420
1.2556	8.0	1168	1.0140	0.5846
0.9216	10.0	1460	0.8558	0.5142
0.7769	12.0	1752	0.7731	0.4643
0.6968	14.0	2044	0.7458	0.4394
0.6813	16.0	2336	0.7549	0.4385
0.5996	18.0	2628	0.7186	0.4128
0.572	20.0	2920	0.7102	0.4126

Framework versions

• Transformers 4.42.3
• Pytorch 2.1.2
• Datasets 2.20.0
• Tokenizers 0.19.1