wav2vec2-base-timit-demo-colab

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: 0.0403
• Wer: 0.0168

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: 64
• eval_batch_size: 16
• 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
4.5738	2.82	500	2.8712	1.0
1.3905	5.65	1000	0.2342	0.2124
0.1868	8.47	1500	0.1023	0.0697
0.0831	11.3	2000	0.0603	0.0339
0.0512	14.12	2500	0.0519	0.0263
0.0363	16.95	3000	0.0478	0.0228
0.0267	19.77	3500	0.0490	0.0228
0.0205	22.6	4000	0.0390	0.0182
0.0163	25.42	4500	0.0418	0.0184
0.0145	28.25	5000	0.0403	0.0168

Framework versions

• Transformers 4.11.3
• Pytorch 1.9.1+cu111
• Datasets 1.13.3
• Tokenizers 0.10.3