dataset:
  name: food101
  img_size: 224
  img_chls: 3
  aug:
    hor_flip_aug: true
    auto_augment: false
    rand_augment: true
  n_class: 101
model:
  name: ViT-B-16
  img_size: ${dataset.img_size}
  patch_size: 16
  img_chls: ${dataset.img_chls}
  n_class: ${dataset.n_class}
  n_layer: 12
  n_heads: 12
  n_embd: 768
  mlp_dim: 3072
  drop_rate: 0.1
  attn_drop_rate: 0.1
  stoch_depth_drop_rate: 0.2
rng_seed: 31415
interactive: true
model_name: ${model.name}-${dataset.name}
n_epochs: 100
save_every_epoch: 20
val_every_epoch: 10
init_from: scratch
device_type: cuda
torch_compile: true
enable_tf32: true
clip_grad_norm_1: true
autocast_dtype: bf16
wandb: true
dataloader:
  batch_size: 256
  num_workers: 4
  pin_memory: true
  drop_last: true
optimizer:
  _target_: torch.optim.AdamW
  lr: 0.0003
  betas:
  - 0.9
  - 0.99
  weight_decay: 0.1
  eps: 1.0e-08
  fused: true
lr_scheduler:
  name: cosine-with-linear-warmup
  min_lr_pct: 0.1
  warmup_pct: 0.1
  decay_step_pct: 0.6
deit:
  enable: false
  teacher_name: null
logging:
  log_dir: ./logs
  wandb:
    enable: ${wandb}
    project: vit