wav2vec2-lv-60-espeak-cv-ft-WCTC-test-phocab-ds-f4

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: 16.9428
• Per: 0.0719

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
1366.0797	0.7194	400	249.5698	1.0
486.5059	1.4388	800	232.5474	1.0
452.9709	2.1583	1200	204.3137	1.0
325.1364	2.8777	1600	70.2269	0.3195
184.4623	3.5971	2000	37.8676	0.1649
142.7856	4.3165	2400	27.1812	0.1256
123.0474	5.0360	2800	22.6812	0.0980
113.6602	5.7554	3200	20.1458	0.0799
106.1549	6.4748	3600	17.9236	0.0741
100.9898	7.1942	4000	17.0826	0.0726
100.329	7.9137	4400	16.9428	0.0719

Framework versions

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