xlsr-wav2vec2-model

This model is a fine-tuned version of facebook/wav2vec2-xls-r-300m on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.7303
• Wer: 0.4284

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: 8
• 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

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
7.3963	2.1834	500	3.4058	1.0
2.9024	4.3668	1000	3.1374	1.0
2.5161	6.5502	1500	1.4918	1.0610
1.1536	8.7336	2000	0.8087	0.6197
0.7102	10.9170	2500	0.6845	0.5452
0.5177	13.1004	3000	0.6658	0.5055
0.3888	15.2838	3500	0.6185	0.4784
0.3311	17.4672	4000	0.6584	0.4690
0.2776	19.6507	4500	0.6956	0.4529
0.24	21.8341	5000	0.6835	0.4510
0.203	24.0175	5500	0.6903	0.4452
0.1945	26.2009	6000	0.7061	0.4332
0.1627	28.3843	6500	0.7303	0.4284

Framework versions

• Transformers 4.40.2
• Pytorch 2.2.1+cu121
• Datasets 2.19.1
• Tokenizers 0.19.1