revised_ft_wav2vec2_base_thirty

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

• Loss: 2.7407
• Wer: 99.8829
• Cer: 98.8775

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: 16
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 2000
• training_steps: 4500

Training results

Training Loss	Epoch	Step	Validation Loss	Wer	Cer
3.5435	3.5088	1000	2.7750	99.9040	98.9988
0.885	7.0175	2000	0.6658	33.7675	14.7845
0.391	10.5263	3000	0.6526	29.5786	13.3104
0.2561	14.0351	4000	0.6522	27.3708	12.2904

Framework versions

• Transformers 4.40.2
• Pytorch 1.12.1+cu116
• Datasets 2.19.1
• Tokenizers 0.19.1