whisper-tiny-en

This model is a fine-tuned version of openai/whisper-tiny on the minds14 dataset. It achieves the following results on the evaluation set:

• Loss: 0.7626
• Wer Ortho: 0.2891
• Wer: 0.2884

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: 32
• eval_batch_size: 16
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 50
• training_steps: 4000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer Ortho	Wer
0.0005	35.71	500	0.6319	0.2684	0.2684
0.0002	71.43	1000	0.6820	0.2709	0.2709
0.0001	107.14	1500	0.7092	0.2740	0.2739
0.0001	142.86	2000	0.7275	0.2854	0.2848
0.0001	178.57	2500	0.7423	0.2885	0.2878
0.0	214.29	3000	0.7531	0.2898	0.2890
0.0	250.0	3500	0.7604	0.2898	0.2890
0.0	285.71	4000	0.7626	0.2891	0.2884

Framework versions

• Transformers 4.39.2
• Pytorch 1.13.0+cu117
• Datasets 2.16.1
• Tokenizers 0.15.1