whisper-large-v2-ec-unfinished

This model is a fine-tuned version of openai/whisper-large-v2 on the wanasash/enwaucymraeg_unfinished default dataset. It achieves the following results on the evaluation set:

• Loss: 0.9406
• Wer: 0.3182

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
0.0037	26.6667	1000	0.7359	0.3172
0.0002	53.3333	2000	0.8618	0.3087
0.0001	80.0	3000	0.9035	0.3134
0.0001	106.6667	4000	0.9301	0.3182
0.0001	133.3333	5000	0.9406	0.3182

Framework versions

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