wav2vec2-final-learning-rate-3e-4

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

• Loss: inf
• Wer: 0.3813
• Cer: 0.0939

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: 32
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 64
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 400
• num_epochs: 20
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer	Cer
1.6352	3.2616	600	inf	0.5060	0.1252
0.386	6.5232	1200	inf	0.4420	0.1102
0.2862	9.7847	1800	inf	0.4021	0.0988
0.1631	13.0436	2400	inf	0.3977	0.0979
0.126	16.3052	3000	inf	0.3861	0.0950
0.0732	19.5668	3600	inf	0.3813	0.0939

Framework versions

• Transformers 4.51.1
• Pytorch 2.1.0+cu118
• Datasets 3.5.0
• Tokenizers 0.21.1