dynann/whisper-small-kh

Whisper Tiny khm - Dynann

This model is a fine-tuned version of openai/whisper-tiny on the Khmer Male V2 dataset. It achieves the following results on the evaluation set:

• eval_loss: 0.3435
• eval_wer: 70.9467
• eval_runtime: 110.956
• eval_samples_per_second: 2.623
• eval_steps_per_second: 0.333
• epoch: 12.2324
• step: 4000

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: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 80
• num_epochs: 20
• mixed_precision_training: Native AMP

Framework versions

• Transformers 4.44.2
• Pytorch 2.6.0+cu124
• Datasets 4.1.1
• Tokenizers 0.19.1