wav2vec2-base-timit-demo-colab_1

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: 0.3233
• Wer: 0.2574

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: 0.0002
• train_batch_size: 16
• 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
4.0949	3.52	500	1.1140	0.7136
0.7584	7.04	1000	0.5312	0.5154
0.4254	10.56	1500	0.4489	0.4401
0.2708	14.08	2000	0.4108	0.3770
0.1855	17.61	2500	0.3881	0.3257
0.139	21.13	3000	0.3666	0.2958
0.1057	24.65	3500	0.3351	0.2748
0.0855	28.17	4000	0.3233	0.2574

Framework versions

• Transformers 4.11.3
• Pytorch 1.11.0+cu113
• Datasets 1.18.3
• Tokenizers 0.10.3