def test_pretrained(self):
kwargs = {
'n_class': self.n_class,
'pretrained_model': self.pretrained_model,
'mean': self.mean,
}
if self.pretrained_model == 'imagenet':
valid = self.n_class in {None, 1000}
if valid:
VGG16(**kwargs)
else:
with self.assertRaises(ValueError):
VGG16(**kwargs)
def main():
args = parse_args()
model = VGG16(pretrained_model='imagenet')
image = cv2.imread(args.image_path)
image_size = (224, 224)
image = cv2.resize(image, image_size)
if args.gpu >= 0:
cuda.get_device_from_id(args.gpu).use()
model.to_gpu()
xp = cuda.cupy
else:
xp = np
x = np.float32(image)
x = x.transpose((2, 0, 1))[np.newaxis, ::-1, :, :]
x -= model.mean
x = xp.asarray(x)
y_grad = xp.zeros((1, 1000), dtype=np.float32)
y_grad[0, args.label] = 1.0
gcam = gradcam.gradcam(model,
x, [model.conv5_3.conv, F.relu],
y_grad=y_grad)
gcam = cuda.to_cpu(gcam[0])
heatmap_image = gradcam.heatmap(gcam, image_size)
cv2.imwrite(args.heatmap_path, heatmap_image)
overlay_image = gradcam.overlay(image, gcam)
cv2.imwrite(args.overlay_path, overlay_image)
def main():
parser = argparse.ArgumentParser(
description='Learning convnet from ILSVRC2012 dataset')
parser.add_argument('val', help='Path to root of the validation dataset')
parser.add_argument(
'--model', choices=('vgg16', 'resnet50', 'resnet101', 'resnet152'))
parser.add_argument('--pretrained_model', default='imagenet')
parser.add_argument('--gpu', type=int, default=-1)
parser.add_argument('--batchsize', type=int, default=32)
parser.add_argument('--crop', choices=('center', '10'), default='center')
parser.add_argument('--resnet_mode', default='he')
args = parser.parse_args()
dataset = DirectoryParsingLabelDataset(args.val)
label_names = directory_parsing_label_names(args.val)
n_class = len(label_names)
iterator = iterators.MultiprocessIterator(
dataset, args.batchsize, repeat=False, shuffle=False,
n_processes=6, shared_mem=300000000)
if args.model == 'vgg16':
extractor = VGG16(n_class, args.pretrained_model)
elif args.model == 'resnet50':
extractor = ResNet50(
n_class, args.pretrained_model, mode=args.resnet_mode)
elif args.model == 'resnet101':
extractor = ResNet101(
n_class, args.pretrained_model, mode=args.resnet_mode)
elif args.model == 'resnet152':
extractor = ResNet152(
n_class, args.pretrained_model, mode=args.resnet_mode)
model = FeaturePredictor(
extractor, crop_size=224, scale_size=256, crop=args.crop)
if args.gpu >= 0:
chainer.cuda.get_device(args.gpu).use()
model.to_gpu()
print('Model has been prepared. Evaluation starts.')
in_values, out_values, rest_values = apply_to_iterator(
model.predict, iterator, hook=ProgressHook(len(dataset)))
del in_values
pred_probs, = out_values
gt_labels, = rest_values
accuracy = F.accuracy(
np.array(list(pred_probs)), np.array(list(gt_labels))).data
print()
print('Top 1 Error {}'.format(1. - accuracy))
def load_vgg16pretrain(model, vggmodel='vgg16convs.mat'):
import chainercv
from chainercv.links import VGG16
cv_model = VGG16(pretrained_model='imagenet')
import torch
model.conv1_1.weight.data[:] = torch.tensor(
cv_model.conv1_1.conv.W.data[:, ::-1].copy())
model.conv1_1.bias.data[:] = torch.tensor(cv_model.conv1_1.conv.b.data)
model.conv1_2.weight.data[:] = torch.tensor(cv_model.conv1_2.conv.W.data)
model.conv1_2.bias.data[:] = torch.tensor(cv_model.conv1_2.conv.b.data)
model.conv2_1.weight.data[:] = torch.tensor(cv_model.conv2_1.conv.W.data)
model.conv2_1.bias.data[:] = torch.tensor(cv_model.conv2_1.conv.b.data)
model.conv2_2.weight.data[:] = torch.tensor(cv_model.conv2_2.conv.W.data)
model.conv2_2.bias.data[:] = torch.tensor(cv_model.conv2_2.conv.b.data)
model.conv3_1.weight.data[:] = torch.tensor(cv_model.conv3_1.conv.W.data)
model.conv3_1.bias.data[:] = torch.tensor(cv_model.conv3_1.conv.b.data)
model.conv3_2.weight.data[:] = torch.tensor(cv_model.conv3_2.conv.W.data)
model.conv3_2.bias.data[:] = torch.tensor(cv_model.conv3_2.conv.b.data)
model.conv3_3.weight.data[:] = torch.tensor(cv_model.conv3_3.conv.W.data)
model.conv3_3.bias.data[:] = torch.tensor(cv_model.conv3_3.conv.b.data)
model.conv4_1.weight.data[:] = torch.tensor(cv_model.conv4_1.conv.W.data)
model.conv4_1.bias.data[:] = torch.tensor(cv_model.conv4_1.conv.b.data)
model.conv4_2.weight.data[:] = torch.tensor(cv_model.conv4_2.conv.W.data)
model.conv4_2.bias.data[:] = torch.tensor(cv_model.conv4_2.conv.b.data)
model.conv4_3.weight.data[:] = torch.tensor(cv_model.conv4_3.conv.W.data)
model.conv4_3.bias.data[:] = torch.tensor(cv_model.conv4_3.conv.b.data)
model.conv5_1.weight.data[:] = torch.tensor(cv_model.conv5_1.conv.W.data)
model.conv5_1.bias.data[:] = torch.tensor(cv_model.conv5_1.conv.b.data)
model.conv5_2.weight.data[:] = torch.tensor(cv_model.conv5_2.conv.W.data)
model.conv5_2.bias.data[:] = torch.tensor(cv_model.conv5_2.conv.b.data)
model.conv5_3.weight.data[:] = torch.tensor(cv_model.conv5_3.conv.W.data)
model.conv5_3.bias.data[:] = torch.tensor(cv_model.conv5_3.conv.b.data)
print('done')
def setUp(self):
self.link = VGG16(n_class=self.n_class,
pretrained_model=None,
initialW=Zero())
self.link.pick = self.pick
def setUp(self):
self.link = VGG16(
n_class=self.n_class, pretrained_model=None,
initialW=Zero())
self.link.feature_names = self.feature_names