eval.yaml

name: InTheWild
description: >
  In-the-Wild audio deepfake dataset (arXiv 2203.16263). Genuine speech and audio
  deepfakes of politicians and public figures collected from publicly available
  sources. 31,779 clips (19,963 bonafide / 11,816 spoof), 16 kHz mono. Binary
  classification: bonafide vs. spoof, scored by EER. A cross-domain
  generalization benchmark.
evaluation_framework: inspect-ai

tasks:
  - id: antispoofing_eval
    config: default
    split: test

    field_spec:
      input: audio
      target: label

    solvers:
      - name: speech_spoof_bench_solver

    scorers:
      - name: speech_spoof_scorer

    metrics:
      - eer_percent