wav2vec2-large-xls-r-1b-malay-colab

This model is a fine-tuned version of facebook/wav2vec2-xls-r-1b on the None dataset. It achieves the following results on the evaluation set:

• Loss: 1.2224
• Wer: 0.3983

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: 0.0003
• train_batch_size: 16
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 32
• optimizer: Use OptimizerNames.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: 500
• num_epochs: 30
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
3.0949	4.2553	400	1.2956	0.7778
1.2835	8.5106	800	1.0625	0.6145
0.8461	12.7660	1200	0.8611	0.4936
0.5795	17.0213	1600	0.9722	0.4649
0.3936	21.2766	2000	1.0780	0.4255
0.2668	25.5319	2400	1.1468	0.4036
0.1717	29.7872	2800	1.2224	0.3983

Framework versions

• Transformers 4.49.0.dev0
• Pytorch 2.5.1+cu121
• Datasets 3.2.1.dev0
• Tokenizers 0.21.0