libri-wav2vec2-xlsr-300m-phoneme-demo

This model is a fine-tuned version of facebook/wav2vec2-xls-r-300m on the None dataset. It achieves the following results on the evaluation set:

• Loss: 17.5394
• Wer: 3.3218
• Cer: 0.9677

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: 1e-05
• train_batch_size: 4
• eval_batch_size: 4
• seed: 42
• gradient_accumulation_steps: 8
• total_train_batch_size: 32
• optimizer: Use adamw_torch 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: 1000
• num_epochs: 20
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer	Cer
No log	5.0	5	17.5401	3.3057	0.9626
No log	10.0	10	17.5399	3.3356	0.9671
No log	15.0	15	17.5399	3.3126	0.9624
No log	20.0	20	17.5394	3.3218	0.9677

Framework versions

• Transformers 4.47.1
• Pytorch 2.5.1+cu124
• Datasets 2.7.1
• Tokenizers 0.21.0