wav2vec2-xls-r-1b_phoneme-timit_english_timit-4k_001

This model is a fine-tuned version of facebook/wav2vec2-xls-r-1b on the timit dataset. It achieves the following results on the evaluation set:

• Loss: 0.4353
• Per: 0.1233

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: 5000
• training_steps: 10000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Per
3.356	1.73	1000	0.4634	0.1786
0.7615	3.46	2000	0.4203	0.1688
0.6183	5.19	3000	0.4097	0.1705
0.6072	6.92	4000	0.4693	0.1798
0.6144	8.65	5000	0.5058	0.1810
0.5871	10.38	6000	0.4735	0.1761
0.5046	12.11	7000	0.4379	0.1654
0.4093	13.84	8000	0.4333	0.1577
0.3299	15.57	9000	0.4281	0.1526
0.2723	17.3	10000	0.4353	0.1499

Framework versions

• Transformers 4.28.1
• Pytorch 2.0.1
• Datasets 2.18.0
• Tokenizers 0.13.3