from datagen import genanchors
print('tensorflow version: {}'.format(tf.__version__))
output_path = 'output'
ishape = [512, 512, 3]
ssize = [128, 128]
asizes = [[32, 32]]
total_classes = 1
resnet_settings = [[8, 8, 32], [24, [2, 2]]]
nsm_iou_threshold = 0.2
nsm_score_threshold = 0.8
nsm_max_output_size = 100
a1box_2dtensor = tf.constant(value=genanchors(isize=ishape[:2],
ssize=ssize,
asizes=asizes),
dtype='float32') # (h1*w1*k1, 4)
a2box_2dtensor = tf.constant(value=genanchors(
isize=[ishape[0] // 2, ishape[1] // 2],
ssize=[ssize[0] // 2, ssize[1] // 2],
asizes=asizes),
dtype='float32') # (h2*w2*k2, 4)
a3box_2dtensor = tf.constant(value=genanchors(
isize=[ishape[0] // 4, ishape[1] // 4],
ssize=[ssize[0] // 4, ssize[1] // 4],
asizes=asizes),
dtype='float32') # (h3*w3*k3, 4)
a4box_2dtensor = tf.constant(value=genanchors(
isize=[ishape[0] // 8, ishape[1] // 8],
ssize=[ssize[0] // 8, ssize[1] // 8],
from datagen import genanchors
print('tensorflow version: {}'.format(tf.__version__))
output_path = 'output'
ishape = [512, 512, 3]
ssize = [128, 128]
asizes = [[32, 32]]
total_classes = 1
resnet_settings = [[8, 8, 32], [24, [2, 2]]]
nsm_iou_threshold = 0.2
nsm_score_threshold = 0.8
nsm_max_output_size = 100
a1box_2dtensor = tf.constant(value=genanchors(isize=ishape[:2], ssize=ssize, asizes=asizes), dtype='float32') # (h1*w1*k1, 4)
a2box_2dtensor = tf.constant(value=genanchors(isize=[ishape[0]//2, ishape[1]//2], ssize=[ssize[0]//2, ssize[1]//2], asizes=asizes), dtype='float32') # (h2*w2*k2, 4)
a3box_2dtensor = tf.constant(value=genanchors(isize=[ishape[0]//4, ishape[1]//4], ssize=[ssize[0]//4, ssize[1]//4], asizes=asizes), dtype='float32') # (h3*w3*k3, 4)
a4box_2dtensor = tf.constant(value=genanchors(isize=[ishape[0]//8, ishape[1]//8], ssize=[ssize[0]//8, ssize[1]//8], asizes=asizes), dtype='float32') # (h4*w4*k4, 4)
abox_2dtensor = tf.concat(values=[a1box_2dtensor, a2box_2dtensor, a3box_2dtensor, a4box_2dtensor], axis=0)
model = build_robust_test_model(
ishape=ishape,
resnet_settings=resnet_settings,
k=len(asizes),
total_classes=total_classes,
abox_2dtensor=abox_2dtensor,
nsm_iou_threshold=nsm_iou_threshold,
nsm_score_threshold=nsm_score_threshold,
nsm_max_output_size=nsm_max_output_size)
combine = True if ishape[0] == 512 else False
ssize = [ishape[0] / 4, ishape[1] / 4]
asizes = [[32, 32]]
total_classes = 1
resnet_settings = [[8, 8, 32], [24, [2, 2]]]
iou_thresholds = [0.3, 0.5]
anchor_sampling = 256
nsm_iou_threshold = 0.2
nsm_score_threshold = 0.8
total_epoches = 1000
nsm_max_output_size = 100
total_train_examples = 4
total_test_examples = 4
abox_2dtensor = tf.constant(value=genanchors(isize=ishape[:2],
ssize=ssize,
asizes=asizes),
dtype='float32') # (h*w*k, 4)
model = build_model(ishape=ishape,
resnet_settings=resnet_settings,
k=len(asizes),
total_classes=total_classes,
net_name=net_name)
# model.summary()
model.load_weights('{}/weights_.h5'.format(output_path), by_name=True)
min_loss = 2**32
max_precision = 0
max_recall = 0
max_precision_recall = 0