def load_model(model_path, encoder_path=None):
model = NeuralNetwork()
model.load(file_name=model_path)
if encoder_path is not None:
encoder = pickle.load(encoder_path)
return model, encoder
else:
return model
def load_model(model_path, encoder=None):
model = NeuralNetwork()
model.load(file_name=model_path)
if encoder is not None:
if type(encoder) == str:
encoder = pickle.load(open(encoder, 'rb'))
return model, encoder
elif type(encoder) == onehot_decoder:
return model, encoder
else:
return model
def pred_func(model: NeuralNetwork, timestep):
global data_q
if len(data_q) > timestep:
data = np.expand_dims(np.asarray(
[data_q.pop() for _ in range(timestep)]),
axis=0)
return model.predict(x=data)
else:
print('Not Enough Data to predict')
return None
p_processed.setXRange(-1, 1)
p_processed.setYRange(-1, 1)
p_processed.setLabel('left', text='Z position (m)')
p_processed.setLabel('bottom', text='Doppler (m/s)')
s_processed = p_processed.plot([], [], pen=None, symbol='o')
# Main loop
detObj = {}
frameData = {}
preprocessed_frameArray = []
# reading RNN model
from keras.models import load_model
if isPredict:
regressive_classifier = NeuralNetwork()
regressive_classifier.load(file_name='trained_models/radar_model/072319_02/regressive_classifier.h5')
onNotOn_ann_classifier = NeuralNetwork()
onNotOn_ann_classifier.load(file_name='F:/config_detection/models/onNotOn_ANN/classifier_080719_2.h5')
onNotOn_encoder = pickle.load(open('F:/config_detection/models/onNotOn_ANN/encoder_080719_2', 'rb'))
rnn_timestep = 100
num_padding = 50
def input_thread(a_list):
input()
interrupt_list.append(True)
class prediction_thread(Thread):
def __init__(self, event):
p_processed.setLabel('bottom', text='Doppler (m/s)')
s_processed = p_processed.plot([], [], pen=None, symbol='o')
# Main loop
detObj = {}
frameData = {}
preprocessed_frameArray = []
frameArray = []
# reading RNN model
from keras.models import load_model
if isPredict:
# regressive_classifier = NeuralNetwork()
# regressive_classifier.load(file_name='trained_models/radar_model/072319_02/regressive_classifier.h5')
onNotOn_ann_classifier = NeuralNetwork()
onNotOn_ann_classifier.load(file_name='F:/config_detection/models/onNotOn_ANN/classifier_080919_2.h5')
onNotOn_encoder = pickle.load(open('F:/config_detection/models/onNotOn_ANN/encoder_080919_2', 'rb'))
rnn_timestep = 100
num_padding = 50
def input_thread(a_list):
input()
interrupt_list.append(True)
class prediction_thread(Thread):
def __init__(self, event):
Thread.__init__(self)
self.stopped = event
def pred_func(model: NeuralNetwork, data):
return model.predict(x=data)