Edit model card

whisper-tiny-finetuned-no-go-kws

This model is a fine-tuned version of openai/whisper-tiny on the Speech Commands[no, go] dataset. It achieves the following results on the evaluation set:

  • Loss: 0.0842
  • Accuracy: 0.9901

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: 5e-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_ratio: 0.1
  • num_epochs: 10
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss Accuracy
0.33 1.0 780 0.0272 0.9938
0.0002 2.0 1560 0.0420 0.9876
0.0001 3.0 2340 0.0487 0.9913
0.0011 4.0 3120 0.0789 0.9802
0.0001 5.0 3900 0.0915 0.9851
0.0014 6.0 4680 0.1017 0.9839
0.0 7.0 5460 0.0993 0.9888
0.0 8.0 6240 0.0694 0.9913
0.0 9.0 7020 0.0760 0.9926
0.0 10.0 7800 0.0842 0.9901

Framework versions

  • Transformers 4.36.0.dev0
  • Pytorch 2.1.0+cu118
  • Datasets 2.15.0
  • Tokenizers 0.15.0
Downloads last month
6
Safetensors
Model size
8.32M params
Tensor type
F32
·

Finetuned from

Dataset used to train Teapack1/whisper-tiny-finetuned-no-go-kws

Evaluation results