# under the License.

import os
import argparse
import logging
logging.basicConfig(level=logging.DEBUG)
from common import find_mxnet, data, fit
from common.util import download_file
import mxnet as mx

if __name__ == '__main__':
    # parse args
    parser = argparse.ArgumentParser(description="train imagenet-1k",
                                     formatter_class=argparse.ArgumentDefaultsHelpFormatter)
    fit.add_fit_args(parser)
    data.add_data_args(parser)
    data.add_data_aug_args(parser)
    # uncomment to set standard augmentation for resnet training
    # data.set_resnet_aug(parser)
    parser.set_defaults(
        # network
        network          = 'resnet',
        num_layers       = 50,
        # data
        num_classes      = 1000,
        num_examples     = 1281167,
        image_shape      = '3,224,224',
        min_random_scale = 1, # if input image has min size k, suggest to use
                              # 256.0/x, e.g. 0.533 for 480
        # train
        num_epochs       = 80,
Beispiel #2
0
import mxnet as mx

def set_imagenet_aug(aug):
    # standard data augmentation setting for imagenet training
    aug.set_defaults(rgb_mean='123.68,116.779,103.939', rgb_std='58.393,57.12,57.375')
    aug.set_defaults(random_crop=0, random_resized_crop=1, random_mirror=1)
    aug.set_defaults(min_random_area=0.08)
    aug.set_defaults(max_random_aspect_ratio=4./3., min_random_aspect_ratio=3./4.)
    aug.set_defaults(brightness=0.4, contrast=0.4, saturation=0.4, pca_noise=0.1)

if __name__ == '__main__':
    # parse args
    parser = argparse.ArgumentParser(description="train imagenet-1k",
                                     formatter_class=argparse.ArgumentDefaultsHelpFormatter)
    fit.add_fit_args(parser)
    data.add_data_args(parser)
    data.add_data_aug_args(parser)
    # uncomment to set standard augmentations for imagenet training
    # set_imagenet_aug(parser)
    parser.set_defaults(
        # network
        network          = 'resnet',
        num_layers       = 50,
        # data
        num_classes      = 1000,
        num_examples     = 1281167,
        image_shape      = '3,224,224',
        min_random_scale = 1, # if input image has min size k, suggest to use
                              # 256.0/x, e.g. 0.533 for 480
        # train
        num_epochs       = 80,