wav2vec2-lv-60-espeak-cv-ft-WCTC-test-phocab-ds-f3

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: 17.7418
• Per: 0.0738

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
1391.6586	0.7194	400	246.4171	1.0
480.0798	1.4388	800	228.4021	1.0
456.8211	2.1583	1200	212.6149	1.0
366.5904	2.8777	1600	86.1591	0.3965
199.1429	3.5971	2000	40.5242	0.1626
150.9429	4.3165	2400	29.5733	0.1324
130.1954	5.0360	2800	24.3561	0.1086
115.769	5.7554	3200	20.7409	0.0864
108.5345	6.4748	3600	19.1609	0.0785
104.571	7.1942	4000	17.9111	0.0714
104.8577	7.9137	4400	17.7418	0.0738

Framework versions

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