whisper-medium-ft-5000

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

• Loss: 0.2673
• Wer: 10.3673

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
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 6

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.5238	1.0	313	0.2389	12.2734
0.1035	2.0	626	0.2360	11.2041
0.0381	3.0	939	0.2349	10.7857
0.0134	4.0	1252	0.2512	10.5532
0.0039	5.0	1565	0.2611	10.2278
0.0013	6.0	1878	0.2673	10.3673

Framework versions

• Transformers 4.33.1
• Pytorch 2.0.1+cu117
• Datasets 2.14.5
• Tokenizers 0.13.3