articuplay1

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

• Loss: 0.2033
• Wer: 13.2533

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 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
• num_epochs: 5
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
2.9702	1.0	289	2.4632	95.9398
0.2224	2.0	578	0.2168	14.6648
0.149	3.0	867	0.1952	13.7123
0.1087	4.0	1156	0.1945	13.2154
0.0763	5.0	1445	0.2033	13.2533

Framework versions

• Transformers 4.51.3
• Pytorch 2.6.0+cu124
• Datasets 3.1.0
• Tokenizers 0.21.0