whisper-small-ar-draligomaa-dataset

This model is a fine-tuned version of openai/whisper-small on the draligomaa-dataset dataset. It achieves the following results on the evaluation set:

• Loss: 0.2903
• Wer Ortho: 19.0868

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: 32
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: constant_with_warmup
• lr_scheduler_warmup_steps: 20
• training_steps: 500
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer Ortho
0.3223	0.7246	100	0.3483	24.8217
0.2086	1.4493	200	0.3080	21.9868
0.1403	2.1739	300	0.2934	20.3997
0.1366	2.8986	400	0.2879	19.9791
0.083	3.6232	500	0.2903	19.0868

Framework versions

• Transformers 4.45.1
• Pytorch 2.4.0
• Datasets 3.0.1
• Tokenizers 0.20.0