import tensorlayer as tl
from tensorlayer.models.imagenet_classes import class_names
tf.logging.set_verbosity(tf.logging.DEBUG)
tl.logging.set_verbosity(tl.logging.DEBUG)
x = tf.placeholder(tf.float32, [None, 224, 224, 3])
// get the whole model
mobilenetv1 = tl.models.MobileNetV1(x)
After Change
img1 = tl.vis.read_image("data/tiger.jpeg")
img1 = tl.prepro.imresize(img1, (224, 224)) / 255
img1 = img1.astype(np.float32)[np.newaxis, ...]
start_time = time.time()
output = mobilenetv1(img1, is_train=False)
prob = tf.nn.softmax(output)[0].numpy()
print(" End time : %.5ss" % (time.time() - start_time))
preds = (np.argsort(prob)[::-1])[0:5]
for p in preds:
