wav2vec-fine_tuned-speech_command2

This model is a fine-tuned version of facebook/wav2vec2-base on the speech_commands dataset. It achieves the following results on the evaluation set:

  • Loss: 0.1040
  • Accuracy: 0.9735

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: 0.0003
  • train_batch_size: 256
  • eval_batch_size: 256
  • seed: 42
  • gradient_accumulation_steps: 4
  • total_train_batch_size: 1024
  • 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

Training results

Training Loss Epoch Step Validation Loss Accuracy
1.3874 1.0 50 0.9633 0.9229
0.5144 2.0 100 0.4398 0.9138
0.3538 3.0 150 0.1688 0.9651
0.2956 4.0 200 0.1622 0.9623
0.2662 5.0 250 0.1425 0.9665
0.2122 6.0 300 0.1301 0.9682
0.1948 7.0 350 0.1232 0.9693
0.1837 8.0 400 0.1116 0.9734
0.1631 9.0 450 0.1041 0.9734
0.1441 10.0 500 0.1040 0.9735

Framework versions

  • Transformers 4.30.2
  • Pytorch 2.0.0
  • Datasets 2.1.0
  • Tokenizers 0.13.3
Downloads last month
47
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Dataset used to train Thamer/wav2vec-fine_tuned-speech_command2