seed: 20260830
deterministic: true
data:
  root: data/raw/community_forensics
  train_manifest: data/manifests/community_forensics_train_v3.csv
  val_manifest: data/manifests/community_forensics_val_unseen_generator.csv
  image_size: 224
  num_workers: 6
  pin_memory: true
  persistent_workers: true
  train_clean_probability: 0.25
  train_single_probability: 0.5
  train_double_probability: 0.25
  validation_slices:
    exact_seen_generator: data/manifests/community_forensics_val_external_exact_seen_generator.csv
    hard_hourglass_exact_seen: data/manifests/community_forensics_val_hard_hourglass_v2_exact_seen.csv
    hard_dfgan_exact_seen: data/manifests/community_forensics_val_hard_dfgan_v2_exact_seen.csv
    hard_galip_exact_seen: data/manifests/community_forensics_val_hard_galip_v2_exact_seen.csv
  external_tests:
    unseen_generator: data/manifests/community_forensics_test_external_unseen_generator.csv
    unseen_generator_expanded: data/manifests/community_forensics_test_external_unseen_generator_v3_expanded.csv
  generator_protocol:
    exact_seen_generator: exact identity seen in train-v3, disjoint evaluation images
    hard_generators: Hourglass, DFGAN, and GALIP remain exact-seen from train-v2 promotion
    unseen_generator: architecture family and exact identity both unseen
    retired_seen_family_test: promoted into train-v2 and forbidden for future evaluation
    train_v3_additions: 1000 GAN and 1000 PixDiff AIGI covering all 12 and 3 exact
      Small generators
    real_counterparts: 1000 GAN-matched and 1000 PixDiff-matched real images; Small
      real-source metadata is N/A
    unseen_generator_expanded: frozen 2k strict-unseen base plus 1k AIGI and 1k source-balanced
      real additions
  format_debias:
    enabled: true
    train_qualities:
    - 70
    - 80
    - 90
    - 95
    eval_quality: 90
    jpeg_subsampling: 2
model:
  experiment: m2
  cnn_backbone: efficientnet_b0
  cnn_pretrained: true
  clip_model: ViT-B-32
  clip_pretrained: laion2b_s34b_b79k
  clip_cache_dir: data/cache/open_clip
  clip_mean:
  - 0.48145466
  - 0.4578275
  - 0.40821073
  clip_std:
  - 0.26862954
  - 0.26130258
  - 0.27577711
  projection_dim: 256
  dropout: 0.2
  forensic:
    patch_size: 56
    patches_per_band: 2
    dct_size: 16
    output_dim: 256
train:
  epochs: 3
  batch_size: 24
  learning_rate: 0.0002
  weight_decay: 0.0001
  warmup_ratio: 0.05
  min_learning_rate_ratio: 0.05
  amp: true
  gradient_accumulation: 2
  grad_clip_norm: 1.0
  log_every_steps: 20
  checkpoint_every_steps: 250
  resume: auto
  lambda_kl: 0.5
  lambda_feature: 0.25
evaluation:
  batch_size: 24
  threshold: auto
  matrix: configs/transforms.yaml
output:
  directory: outputs/community_forensics_v3/m2