supervised vith16 448 yaml

configs/supervised_wilds_vith16-448_ep300-lp.yaml

@@ -0,0 +1,60 @@
+meta:
+  model_name: vit_huge
+  embed_dim: 1280
+  load_checkpoint: true
+  checkpoint_folder: experiment_logs/imagenet-vith16.448/
+  read_checkpoint: IN1K-vit.h.16-448px-300e.pth.tar
+  use_bfloat16: true
+  num_classes: 182
+  representation_type: last_avgpool
+  head_type: linear
+
+data:
+  batch_size: 1024
+  root_path: ./wilds_data
+  num_workers: 10
+  pin_mem: true
+  crop_size: 448
+  crop_scale: [1.0, 1.0]
+  use_random_resized_crop: false
+  use_horizontal_flip: false
+  use_color_distortion: false
+  color_jitter_strength: 0.0
+  use_gaussian_blur: false
+
+mask:
+  patch_size: 16
+
+optimization:
+  optimizer: adamw # 'adamw', 'sgd', or 'lars'
+  freeze_weights: true # true for linear probing, false for full fine-tuning
+  epochs: 300
+  lr: 0.01
+  weight_decay: 5.0e-4
+  lr_schedule: cosine
+  step_milestones: [15, 30, 45]
+  step_gamma: 0.1
+  start_lr: 0.0
+  final_lr: 0.0
+  warmup: 0
+  momentum: 0.99
+  lars_eta: 0.001
+  lars_eps: 1.0e-8
+  lars_exclude_bias_and_norm: true
+  ipe_scale: 1.0
+  # Sweep suggestions (manual edits):
+  # - lr: 0.01 | 0.05 | 0.001
+  # - weight_decay: 5.0e-4 | 0.0
+  early_stopping:
+    enabled: true
+    patience: 15
+    min_delta: 1.0e-4
+    min_epochs: 15
+    restore_best_weights: true
+
+validation:
+  eval_every: 1
+
+logging:
+  write_tag: linear_probe
+  auto_folder: true