wav2vec2-large-xlsr-53-demo-colab

This model is a fine-tuned version of facebook/wav2vec2-large-xlsr-53 on the common_voice dataset. It achieves the following results on the evaluation set:

• Loss: 0.4058
• Wer: 0.4799

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: 0.0003
• train_batch_size: 16
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 32
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• num_epochs: 30
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
5.0985	4.26	400	2.2067	1.0201
0.7695	8.51	800	0.4558	0.6311
0.3145	12.77	1200	0.3946	0.5395
0.2208	17.02	1600	0.4187	0.5259
0.1656	21.28	2000	0.4176	0.5038
0.1311	25.53	2400	0.3971	0.4842
0.1195	29.79	2800	0.4058	0.4799

Framework versions

• Transformers 4.28.1
• Pytorch 2.0.0+cu118
• Datasets 2.12.0
• Tokenizers 0.13.3