wav2vec2-xlsr-53-ft-btb-cv-s99-cy

This model is a fine-tuned version of facebook/wav2vec2-large-xlsr-53 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.5434
• Wer: 0.4161

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: 0.0003
• train_batch_size: 8
• eval_batch_size: 64
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 800
• training_steps: 8000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
4.9477	0.0337	500	1.7842	0.9396
1.035	0.0675	1000	1.2386	0.7851
0.8162	0.1012	1500	1.0027	0.7022
0.702	0.1349	2000	0.9614	0.6609
0.6574	0.1687	2500	0.8651	0.6165
0.602	0.2024	3000	0.7874	0.5728
0.5547	0.2361	3500	0.7803	0.5767
0.5325	0.2699	4000	0.7444	0.5494
0.5074	0.3036	4500	0.6984	0.5280
0.4755	0.3373	5000	0.6562	0.4964
0.4582	0.3711	5500	0.6308	0.4779
0.4177	0.4048	6000	0.6201	0.4672
0.3892	0.4385	6500	0.5882	0.4502
0.3694	0.4723	7000	0.5677	0.4295
0.3437	0.5060	7500	0.5474	0.4185
0.3481	0.5397	8000	0.5434	0.4161

Framework versions

• Transformers 4.44.0
• Pytorch 2.4.0+cu121
• Datasets 2.21.0
• Tokenizers 0.19.1