Hemg's picture
Model save
fdefac5 verified
|
raw
history blame
2.68 kB
metadata
license: apache-2.0
base_model: facebook/wav2vec2-base
tags:
  - generated_from_trainer
datasets:
  - audiofolder
metrics:
  - accuracy
model-index:
  - name: deeepfake-audio-Recognition-ttoo
    results:
      - task:
          name: Audio Classification
          type: audio-classification
        dataset:
          name: audiofolder
          type: audiofolder
          config: default
          split: train
          args: default
        metrics:
          - name: Accuracy
            type: accuracy
            value: 0.9696969696969697

deeepfake-audio-Recognition-ttoo

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

  • Loss: 0.1858
  • Accuracy: 0.9697

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: 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: 0.01
  • num_epochs: 16

Training results

Training Loss Epoch Step Validation Loss Accuracy
0.6477 1.0 33 0.5411 0.8182
0.5099 2.0 66 0.4458 0.7879
0.3855 3.0 99 0.5405 0.7727
0.2943 4.0 132 0.2268 0.9394
0.2818 5.0 165 0.2283 0.9091
0.2378 6.0 198 0.1955 0.9394
0.1321 7.0 231 0.2335 0.9394
0.1688 8.0 264 0.2009 0.9545
0.07 9.0 297 0.2629 0.9242
0.0413 10.0 330 0.2156 0.9545
0.0229 11.0 363 0.3189 0.9394
0.0062 12.0 396 0.3850 0.9242
0.0159 13.0 429 0.2462 0.9394
0.0179 14.0 462 0.1904 0.9697
0.0059 15.0 495 0.1844 0.9697
0.0248 16.0 528 0.1858 0.9697

Framework versions

  • Transformers 4.39.2
  • Pytorch 2.2.1+cu121
  • Datasets 2.18.0
  • Tokenizers 0.15.2