Alsman68/whisper-capstone-aviation-audio-trained

capstone-whisper-our-data-only

This model is a fine-tuned version of openai/whisper-medium on the Simulated Aviation Audio For Capstone dataset. It achieves the following results on the evaluation set:

• Loss: 1.9014
• Wer: 52.4476

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: 8
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• training_steps: 4000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.0001	71.4286	1000	1.6081	53.1469
0.0001	142.8571	2000	1.7786	52.6383
0.0	214.2857	3000	1.8591	52.3204
0.0	285.7143	4000	1.9014	52.4476

Framework versions

• Transformers 4.50.1
• Pytorch 2.6.0+cu124
• Datasets 3.4.1
• Tokenizers 0.21.1