checkpointing_steps: null
dataset_name: cifar10
gradient_accumulation_steps: 1
hub_model_id: null
hub_token: null
ignore_mismatched_sizes: false
image_column_name: img
label_column_name: label
learning_rate: 0.001
lr_scheduler_type: cosine
max_eval_samples: null
max_train_samples: null
max_train_steps: 64000
model_name_or_path: MODELS/best_0409_2
num_train_epochs: 193
num_warmup_steps: 6400
num_workers: 32
output_dir: OUTPUTS/best_0409_2--lr0.001--TwoAug
per_device_eval_batch_size: 8
per_device_train_batch_size: 64
push_to_hub: false
report_to: tensorboard
resume_from_checkpoint: null
seed: 42
train_dir: null
train_val_split: 0.15
trust_remote_code: false
validation_dir: null
weight_decay: 0.0
with_tracking: true