wav2vec2-lv-60-espeak-cv-ft-WCTC-test-phocab-ds-f9

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: 20.0792
• Per: 0.0809

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
1369.6837	0.7194	400	252.4306	1.0
479.1117	1.4388	800	235.5406	1.0
452.5693	2.1583	1200	212.1770	1.0
339.6972	2.8777	1600	84.5372	0.3870
198.684	3.5971	2000	47.5357	0.2201
151.5883	4.3165	2400	33.2956	0.1425
128.7607	5.0360	2800	27.4586	0.1088
116.4424	5.7554	3200	23.3936	0.0852
108.6601	6.4748	3600	21.6363	0.0801
104.7405	7.1942	4000	20.3341	0.0784
103.1989	7.9137	4400	20.0792	0.0809

Framework versions

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