meta:
  model_name: vit_huge
  embed_dim: 1280
  load_checkpoint: true
  read_checkpoint: jepa-latest.pth.tar
  use_bfloat16: true
  num_classes: 182

data:
  batch_size: 128
  root_path: ./wilds_data
  num_workers: 10
  pin_mem: true
  crop_size: 224

optimization:
  optimizer: adamw # 'adamw' or 'sgd'
  freeze_weights: true # true for linear probing, false for full fine-tuning
  epochs: 200 # can be set higher if early_stopping
  lr: 0.001
  weight_decay: 0.05
  use_cosine_schedule: false
  start_lr: 0.0001
  final_lr: 1.0e-06
  warmup: 5
  ipe_scale: 1.0
  early_stopping:
    enabled: true
    patience: 6
    min_delta: 0.0
    min_epochs: 5
    restore_best_weights: true

validation:
  eval_every: 1

logging:
  folder: experiment_logs/supervised-vith14.224-bs.128-ep.300/
  write_tag: linear_probe