from __future__ import absolute_import
import keras
from keras import backend as K
from keras.applications.vgg16 import VGG16
from keras.utils.data_utils import get_file
WEIGHTS_PATH = 'https://s3-us-west-2.amazonaws.com/kaggleglm/vgg16_hybrid1365.h5'
WEIGHTS_PATH_NO_TOP = 'https://s3-us-west-2.amazonaws.com/kaggleglm/vgg16_hybrid1365_notop.h5'
def preprocess_input(x, mode='tf'):
x[:, :, 0] -= 104.006
x[:, :, 1] -= 116.669
x[:, :, 2] -= 122.679
# 'RGB'->'BGR'
x = x[:, :, ::-1]
return x
def VGG16PlacesHybrid1365(include_top=True,
weights='imagenet',
classes=1365,
input_shape=(128,128,3),
**kwargs):
model = VGG16(
weights=None, classes=classes, input_shape=input_shape, include_top=include_top, **kwargs)
if weights:
if include_top:
weights_path = get_file(
'vgg16_hybrid1365.h5',
WEIGHTS_PATH,
cache_subdir='models',
file_hash='3ddd2396e124c93143b9bd5d1835e10e')
model.load_weights(weights_path)
else:
weights_path = get_file(
'vgg16_hybrid1365_notop.h5',
WEIGHTS_PATH_NO_TOP,
cache_subdir='models',
file_hash='696badfd31f1195212e3501c8edfc4e4')
model.load_weights(weights_path)
return model
