b32-wav2vec2-large-xls-r-romansh-colab

This model is a fine-tuned version of facebook/wav2vec2-xls-r-300m on the common_voice_13_0 dataset. It achieves the following results on the evaluation set:

• Loss: 0.4636
• Wer: 0.4585

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: 4
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 8
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• num_epochs: 30
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
8.8775	3.05	400	3.2335	1.0
3.0144	6.11	800	2.9346	1.0
2.919	9.16	1200	2.8833	0.9988
1.8698	12.21	1600	0.8435	0.6490
0.6704	15.27	2000	0.5729	0.5249
0.448	18.32	2400	0.4981	0.4823
0.3501	21.37	2800	0.4763	0.4662
0.2999	24.43	3200	0.4610	0.4567
0.2773	27.48	3600	0.4636	0.4585

Framework versions

• Transformers 4.26.0
• Pytorch 2.0.1+cu118
• Datasets 2.14.4
• Tokenizers 0.13.3