w2v2-base-pretrained_lr1e-4_at0.9_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: 6.9410
• Wer: 1.0571

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
17.8057	6.1	250	4.5611	1.0
3.2734	12.2	500	3.7357	1.0
2.8657	18.29	750	2.7315	1.0
0.3941	24.39	1000	4.7924	1.0529
0.1195	30.49	1250	5.4586	1.0507
0.0721	36.59	1500	6.5058	1.0588
0.0472	42.68	1750	6.8300	1.0614
0.0308	48.78	2000	6.8007	1.0597
0.0247	54.88	2250	6.9410	1.0571

Framework versions

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