model

This model is a fine-tuned version of openai/whisper-medium on the None dataset. It achieves the following results on the evaluation set:

• Loss: 5.0046
• Wer: 116.8945

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: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• training_steps: 8000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
1.5232	4.95	1000	3.6227	127.2695
0.0538	9.9	2000	4.3761	125.3417
0.0166	14.85	3000	4.6306	114.6863
0.0008	19.8	4000	4.7625	116.3687
0.0022	24.75	5000	4.9290	116.0182
0.0002	29.7	6000	4.9100	118.2264
0.0001	34.65	7000	4.9886	116.5089
0.0001	39.6	8000	5.0046	116.8945

Framework versions

• Transformers 4.25.0.dev0
• Pytorch 1.12.0+cu116
• Datasets 2.6.1
• Tokenizers 0.13.1