golesheed/whisper-non-native-adult-5-dutch

Whisper Large V2

This model is a fine-tuned version of openai/whisper-large-v2 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.4256
• Wer: 14.5497

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: 3e-05
• train_batch_size: 16
• 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: 20
• num_epochs: 5

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.773	0.55	30	0.4063	23.2487
0.3731	1.09	60	0.3667	14.0416
0.2124	1.64	90	0.3515	18.8453
0.176	2.18	120	0.3743	14.8114
0.0988	2.73	150	0.3732	13.2256
0.0684	3.27	180	0.3910	14.3957
0.0413	3.82	210	0.3921	14.9962
0.0252	4.36	240	0.4235	14.3957
0.0151	4.91	270	0.4256	14.5497

Framework versions

• Transformers 4.38.0.dev0
• Pytorch 2.1.0+cu121
• Datasets 2.14.6
• Tokenizers 0.15.0