myst-wav2vec2-xlsr-300m-03012025

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: 24.3787
• Wer: 1.9278
• Cer: 1.8703

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: 6
• total_train_batch_size: 24
• 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: 10
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer	Cer
No log	1.2143	5	24.3873	1.9589	1.9099
No log	2.4286	10	24.3874	1.9639	1.9189
No log	3.6429	15	24.3867	1.9661	1.9178
No log	4.8571	20	24.3859	1.9517	1.8992
No log	6.2143	25	24.3852	1.9560	1.9067
No log	7.4286	30	24.3839	1.9466	1.8967
No log	8.6429	35	24.3815	1.9358	1.8842
No log	9.8571	40	24.3787	1.9278	1.8703

Framework versions

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