tun_wav2vec7

This model was trained from scratch on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 1.2586
• Wer: 0.5945
• Cer: 0.1592

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: 10
• 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: 70
• num_epochs: 100

Training results

Training Loss	Epoch	Step	Validation Loss	Wer	Cer
2.0634	5.4545	300	1.3331	0.9794	0.3126
0.7466	10.9091	600	1.1736	0.8522	0.2637
0.496	16.3636	900	1.0844	0.8076	0.2344
0.3864	21.8182	1200	1.5553	0.7766	0.2436
0.2908	27.2727	1500	1.1381	0.7045	0.2089
0.2241	32.7273	1800	1.4912	0.7182	0.2236
0.175	38.1818	2100	1.1890	0.6426	0.1910
0.1483	43.6364	2400	1.1582	0.6323	0.1667
0.1237	49.0909	2700	1.2264	0.6357	0.1722
0.1676	54.5455	3000	1.1899	0.6632	0.1726
0.1437	60.0	3300	1.2421	0.6323	0.1793
0.1261	65.4545	3600	1.2693	0.6117	0.1663
0.0787	70.9091	3900	1.2301	0.6014	0.1601
0.0952	76.3636	4200	1.2063	0.6186	0.1659
0.1385	81.8182	4500	1.2732	0.6082	0.1642
0.0659	87.2727	4800	1.3254	0.6048	0.1692
0.058	92.7273	5100	1.2216	0.5979	0.1575
0.1454	98.1818	5400	1.2586	0.5945	0.1592

Framework versions

• Transformers 4.42.3
• Pytorch 2.1.2
• Datasets 2.20.0
• Tokenizers 0.19.1