revised_ft_wav2vec2_base_six_1000

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: 1.1179
• Wer: 57.9258
• Cer: 25.3159

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
3.2301	18.5185	1000	1.1192	57.1980	25.0173
0.4544	37.0370	2000	1.1278	42.9675	19.9377
0.1702	55.5556	3000	1.2664	40.7060	18.8307
0.0902	74.0741	4000	1.2946	39.2168	18.2196

Framework versions

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