def load(self, path_to_model, mode="cpu", verbose=0):
if path_to_model == "latest":
model_info = download_latest_model("TextDetector",
self.get_classname(),
mode=mode)
path_to_model = model_info["path"]
self.MODEL = load_model(path_to_model, compile=False)
net_inp = self.MODEL.get_layer(
name='{}'.format(self.MODEL.layers[0].name)).input
net_out = self.MODEL.get_layer(
name='{}'.format(self.MODEL.layers[-1].name)).output
self.MODEL = Model(inputs=net_inp, outputs=net_out)
if verbose:
self.MODEL.summary()
return self.MODEL
def load(self, path_to_model, mode="cpu", verbose = 0):
if path_to_model =="latest":
model_info = download_latest_model("TextDetector", self.get_classname(), mode=mode)
path_to_model = model_info["path"]
self.MODEL = load_model(path_to_model, compile=False)
net_inp = self.MODEL.get_layer(name='the_input_{}'.format(type(self).__name__)).input
net_out = self.MODEL.get_layer(name='softmax_{}'.format(type(self).__name__)).output
self.MODEL = Model(input=net_inp, output=net_out)
if verbose:
self.MODEL.summary()
# the loss calc occurs elsewhere, so use a dummy lambda func for the loss
#sgd = SGD(lr=0.02, decay=1e-6, momentum=0.9, nesterov=True, clipnorm=5)
#self.MODEL.compile(loss={'ctc': lambda y_true, y_pred: y_pred}, optimizer=sgd)
return self.MODEL