wav2vec2-lv-60-espeak-cv-ft-WCTC-test-phocab-ds-f6

This model is a fine-tuned version of facebook/wav2vec2-lv-60-espeak-cv-ft on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 18.1576
• Per: 0.0722

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: 3e-05
• train_batch_size: 16
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 32
• optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• num_epochs: 8
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Per
1370.168	0.7194	400	249.4827	1.0
482.3394	1.4388	800	233.3257	1.0
458.4452	2.1583	1200	222.9946	1.0
379.1232	2.8777	1600	95.4084	0.4427
205.3075	3.5971	2000	44.5467	0.1819
154.0601	4.3165	2400	31.3072	0.1280
129.7028	5.0360	2800	24.0792	0.1001
117.4587	5.7554	3200	21.0931	0.0818
108.5555	6.4748	3600	19.1006	0.0760
105.3729	7.1942	4000	18.5839	0.0770
103.8013	7.9137	4400	18.1576	0.0722

Framework versions

• Transformers 4.57.6
• Pytorch 2.9.1+cu128
• Datasets 4.5.0
• Tokenizers 0.22.2