whisper-hi-1

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

• Loss: 0.7841
• Wer: 52.1739

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: 4
• eval_batch_size: 4
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 8
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• training_steps: 1000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.0001	30.77	200	0.7062	54.3478
0.0	61.54	400	0.7435	56.5217
0.0	92.31	600	0.7661	54.3478
0.0	123.08	800	0.7792	54.3478
0.0	153.85	1000	0.7841	52.1739

Framework versions

• Transformers 4.39.2
• Pytorch 2.2.0+cu121
• Datasets 2.18.0
• Tokenizers 0.15.2