| import torch.nn.functional as F | |
| # Seed | |
| SEED = 1 | |
| # Dataset | |
| CLASSES = ( | |
| "Airplane", | |
| "Automobile", | |
| "Bird", | |
| "Cat", | |
| "Deer", | |
| "Dog", | |
| "Frog", | |
| "Horse", | |
| "Ship", | |
| "Truck", | |
| ) | |
| SHUFFLE = True | |
| DATA_DIR = "../data" | |
| NUM_WORKERS = 4 | |
| PIN_MEMORY = True | |
| # Training Hyperparameters | |
| CRITERION = F.cross_entropy | |
| INPUT_SIZE = (3, 32, 32) | |
| NUM_CLASSES = 10 | |
| LEARNING_RATE = 0.001 | |
| WEIGHT_DECAY = 1e-4 | |
| BATCH_SIZE = 512 | |
| NUM_EPOCHS = 24 | |
| DROPOUT_PERCENTAGE = 0.05 | |
| LAYER_NORM = "bn" # Batch Normalization | |
| # OPTIMIZER & SCHEDULER | |
| LRFINDER_END_LR = 0.1 | |
| LRFINDER_NUM_ITERATIONS = 50 | |
| LRFINDER_STEP_MODE = "exp" | |
| OCLR_DIV_FACTOR = 100 | |
| OCLR_FINAL_DIV_FACTOR = 100 | |
| OCLR_THREE_PHASE = False | |
| OCLR_ANNEAL_STRATEGY = "linear" | |
| # Compute Related | |
| ACCELERATOR = "cpu" | |
| PRECISION = 32 | |
| # Store | |
| TRAINING_STAT_STORE = "Store/training_stats.csv" | |
| MODEL_SAVE_PATH = "Store/model.pth" | |
| PRED_STORE_PATH = "Store/pred_store.pth" | |
| EXAMPLE_IMG_PATH = "Store/examples/" | |
| # Visualization | |
| NORM_CONF_MAT = True | |