| 29 | #os.environ["WANDB_MODE"] = "offline" |
| 30 | |
| 31 | def get_args_parser(): |
| 32 | config_parser = parser = argparse.ArgumentParser( |
| 33 | description="Training Config", add_help=False |
| 34 | ) |
| 35 | parser.add_argument("-c", "--config",default="imagenet.yml", type=str, metavar="FILE", help="YAML config file specifying default arguments",) |
| 36 | |
| 37 | # important params |
| 38 | parser = argparse.ArgumentParser('MAE fine-tuning for image classification', add_help=False) |
| 39 | parser.add_argument('--batch_size', default=64, type=int, |
| 40 | help='Batch size per GPU (effective batch size is batch_size * accum_iter * # gpus') |
| 41 | parser.add_argument('--epochs', default=200, type=int) |
| 42 | parser.add_argument('--accum_iter', default=1, type=int, |
| 43 | help='Accumulate gradient iterations (for increasing the effective batch size under memory constraints)') |
| 44 | parser.add_argument('--finetune', default='', |
| 45 | help='finetune from checkpoint') |
| 46 | parser.add_argument('--data_path', default='', type=str, |
| 47 | help='dataset path') |
| 48 | parser.add_argument('--exp', default='', type=str, |
| 49 | help='expreiment name') |
| 50 | # Model parameters |
| 51 | parser.add_argument('--model', default='maxformer_10_384', type=str, metavar='MODEL', |
| 52 | help='Name of model to train') |
| 53 | parser.add_argument('--time_step', default=4, type=int, |
| 54 | help='images input size') |
| 55 | parser.add_argument('--input_size', default=224, type=int, |
| 56 | help='images input size') |
| 57 | |
| 58 | parser.add_argument('--drop_path', type=float, default=0.1, metavar='PCT', |
| 59 | help='Drop path rate (default: 0.1)') |
| 60 | |
| 61 | # Optimizer parameters |
| 62 | parser.add_argument('--clip_grad', type=float, default=None, metavar='NORM', |
| 63 | help='Clip gradient norm (default: None, no clipping)') |
| 64 | parser.add_argument('--weight_decay', type=float, default=0.05, |
| 65 | help='weight decay (default: 0.05)') |
| 66 | |
| 67 | parser.add_argument('--lr', type=float, default=None, metavar='LR', |
| 68 | help='learning rate (absolute lr)') |
| 69 | parser.add_argument('--blr', type=float, default=12e-4, metavar='LR', #10 |
| 70 | help='base learning rate: absolute_lr = base_lr * total_batch_size / 256') |
| 71 | parser.add_argument('--layer_decay', type=float, default=1.0, |
| 72 | help='layer-wise lr decay from ELECTRA/BEiT') |
| 73 | |
| 74 | parser.add_argument('--min_lr', type=float, default=5.0e-6, metavar='LR', |
| 75 | help='lower lr bound for cyclic schedulers that hit 0') |
| 76 | |
| 77 | parser.add_argument('--warmup_epochs', type=int, default=5, metavar='N', |
| 78 | help='epochs to warmup LR') |
| 79 | |
| 80 | # Augmentation parameters |
| 81 | parser.add_argument('--color_jitter', type=float, default=0.4, metavar='PCT', |
| 82 | help='Color jitter factor (enabled only when not using Auto/RandAug)')#None |
| 83 | parser.add_argument('--aa', type=str, default='rand-m9-mstd0.5-inc1', metavar='NAME', |
| 84 | help='Use AutoAugment policy. "v0" or "original". " + "(default: rand-m9-mstd0.5-inc1)'), #rand-m9-mstd0.5-inc1 |
| 85 | parser.add_argument('--smoothing', type=float, default=0.1, |
| 86 | help='Label smoothing (default: 0.1)') |
| 87 | |
| 88 | # * Random Erase params |