'export_dir': None,
'precision': 'fp16',
'momentum': 0.9,
'learning_rate_init': 1.0,
'learning_rate_power': 2.0,
'weight_decay': 1e-4,
'loss_scale': 2048.0,
'larc_eta': 0.003,
'larc_mode': 'clip',
'num_iter': 90,
'iter_unit': 'epoch',
'checkpoint_secs': None,
'display_every': 10,
}
args, _ = nvutils.parse_cmdline(default_args)
def inception_v3(inputs, training=False):
"""Google's Inception v3 model
https://arxiv.org/abs/1512.00567
"""
def inception_v3_a(builder, x, n):
cols = [[('conv2d', 64, 1, 1, 'SAME')],
[('conv2d', 48, 1, 1, 'SAME'), ('conv2d', 64, 5, 1, 'SAME')],
[('conv2d', 64, 1, 1, 'SAME'), ('conv2d', 96, 3, 1, 'SAME'),
('conv2d', 96, 3, 1, 'SAME')],
[('apool2d', 3, 1, 'SAME'), ('conv2d', n, 1, 1, 'SAME')]]
return builder.inception_module(x, 'incept_v3_a', cols)
def inception_v3_b(builder, x):
'num_iter': 90,
'iter_unit': 'epoch',
'checkpoint_secs': None,
'display_every': 10,
}
formatter = argparse.ArgumentDefaultsHelpFormatter
parser = argparse.ArgumentParser(formatter_class=formatter)
parser.add_argument('--layers',
default=50,
type=int,
required=True,
choices=[11, 13, 16, 19],
help="""Number of VGG layers.""")
args, flags = nvutils.parse_cmdline(default_args, parser)
def inference_vgg_impl(builder, inputs, layer_counts):
x = inputs
for _ in range(layer_counts[0]):
x = builder.conv2d(x, 64, 3, 1, 'SAME')
x = builder.max_pooling2d(x, 2, 2)
for _ in range(layer_counts[1]):
x = builder.conv2d(x, 128, 3, 1, 'SAME')
x = builder.max_pooling2d(x, 2, 2)
for _ in range(layer_counts[2]):
x = builder.conv2d(x, 256, 3, 1, 'SAME')
x = builder.max_pooling2d(x, 2, 2)
for _ in range(layer_counts[3]):
x = builder.conv2d(x, 512, 3, 1, 'SAME')