finetune_wav2vec2_base_thirty

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: 1.3014
• Wer: 68.3822
• Cer: 31.2142

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
3.4046	2.8090	1000	1.3290	68.5064	31.3589
0.7428	5.6180	2000	0.6691	34.7663	15.3556
0.4195	8.4270	3000	0.6144	29.5906	13.1851
0.2851	11.2360	4000	0.6656	27.3526	12.3476

Framework versions

• Transformers 4.40.2
• Pytorch 2.1.1+cu121
• Datasets 2.19.1
• Tokenizers 0.19.1