whisper-large-v3-ec

This model is a fine-tuned version of openai/whisper-large-v3 on the wanasash/enwaucymraeg default dataset. It achieves the following results on the evaluation set:

• Loss: 0.4733
• Wer: 0.2137

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: 16
• eval_batch_size: 16
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 32
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• training_steps: 5000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.0091	13.6054	1000	0.4027	0.2189
0.0026	27.2109	2000	0.4282	0.2260
0.0002	40.8163	3000	0.4444	0.2178
0.0001	54.4218	4000	0.4667	0.2160
0.0001	68.0272	5000	0.4733	0.2137

Framework versions

• Transformers 4.44.0
• Pytorch 2.4.0+cu121
• Datasets 2.20.0
• Tokenizers 0.19.1