Skip to content

Commit

Permalink
change hyperparm for shadow model
Browse files Browse the repository at this point in the history
  • Loading branch information
snoop2head committed Jul 11, 2022
1 parent a629414 commit ce907ad
Showing 1 changed file with 2 additions and 2 deletions.
4 changes: 2 additions & 2 deletions config.yaml
Original file line number Diff line number Diff line change
Expand Up @@ -9,10 +9,10 @@ CFG:
target_train_size: 0.8 # fraction to divide CIFAR test dataset for target model training
seed: 42
num_classes: 10 # CIFAR10 vs CIFAR100 selection
learning_rate: 0.0005 # Efficientnet Learning rate
learning_rate: 0.005 # Efficientnet Learning rate
weight_decay: 0.00005 # default lr: wd ratio is 0.1(https://github.com/clovaai/AdamP), but using 0.01 for small dataset
input_resolution: 32 # 32 x 32 is cifar10 and cifar100 original image resolution
train_batch_size: 512
train_batch_size: 1024
val_batch_size: 512
logging_steps: 1000
save_path: ./ckpt
Expand Down

0 comments on commit ce907ad

Please sign in to comment.