whisper-base-aug-30jan

This model is a fine-tuned version of PhanithLIM/xwhisper-kh-base on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.2004
• Wer: 86.8567

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: 5e-06
• train_batch_size: 16
• eval_batch_size: 8
• seed: 42
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: constant
• lr_scheduler_warmup_steps: 1000
• num_epochs: 8
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.2977	1.0	381	0.2424	91.3647
0.2423	2.0	762	0.2258	89.7310
0.2164	3.0	1143	0.2156	88.2938
0.1977	4.0	1524	0.2102	87.7533
0.1821	5.0	1905	0.2055	86.5250
0.1683	6.0	2286	0.2029	88.2815
0.1572	7.0	2667	0.2008	88.9203
0.1466	8.0	3048	0.2004	86.8567

Framework versions

• Transformers 4.48.1
• Pytorch 2.5.1+cu121
• Datasets 3.2.0
• Tokenizers 0.21.0