w2v2-base-pretrained_lr1e-4_at0.3_da1

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

• Loss: 6.0832
• Wer: 1.0503

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: 32
• 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: 60
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
18.217	3.97	250	4.5592	1.0
3.272	7.94	500	3.7936	1.0
2.4362	11.9	750	2.3296	1.0085
0.4638	15.87	1000	3.0200	1.0264
0.2526	19.84	1250	3.8211	1.0558
0.1458	23.81	1500	4.3826	1.0439
0.0912	27.78	1750	4.6974	1.0401
0.0637	31.75	2000	5.7380	1.0469
0.0492	35.71	2250	5.6284	1.0520
0.0413	39.68	2500	5.8445	1.0362
0.031	43.65	2750	5.8013	1.0473
0.0248	47.62	3000	6.1615	1.0456
0.0228	51.59	3250	5.9906	1.0439
0.016	55.56	3500	6.1408	1.0494
0.0159	59.52	3750	6.0832	1.0503

Framework versions

• Transformers 4.35.0
• Pytorch 2.0.0
• Datasets 2.14.6
• Tokenizers 0.14.1