Model_P_Wav2Vec2_Versi2

This model is a fine-tuned version of facebook/wav2vec2-large-xlsr-53 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 2.7372
• Wer: 0.5827
• Cer: 0.2585

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

Training results

Training Loss	Epoch	Step	Validation Loss	Wer	Cer
3.6262	3.49	400	2.1100	0.9084	0.3800
0.6747	6.99	800	2.2525	0.7047	0.2942
0.4115	10.48	1200	2.2371	0.6452	0.2875
0.2946	13.97	1600	2.4240	0.6279	0.2705
0.2252	17.47	2000	2.5509	0.6429	0.2759
0.1822	20.96	2400	2.4377	0.6095	0.2692
0.1318	24.45	2800	2.6908	0.5807	0.2549
0.1001	27.95	3200	2.7372	0.5827	0.2585

Framework versions

• Transformers 4.31.0
• Pytorch 2.0.1+cu117
• Datasets 2.14.4
• Tokenizers 0.13.3