def main():
parser = argparse.ArgumentParser(
description='Model Runner for team rwightman')
parser.add_argument('bagfile', type=str, help='Path to ROS bag')
parser.add_argument('--alpha',
type=float,
default=0.1,
help='Path to the metagraph path')
parser.add_argument('--graph_path',
type=str,
help='Path to the self contained graph def')
parser.add_argument('--metagraph_path',
type=str,
help='Path to the metagraph path')
parser.add_argument('--checkpoint_path',
type=str,
help='Path to the checkpoint path')
parser.add_argument('--debug_print',
dest='debug_print',
action='store_true',
help='Debug print of predicted steering commands')
args = parser.parse_args()
def get_model():
model = RwightmanModel(alpha=args.alpha,
graph_path=args.graph_path,
metagraph_path=args.metagraph_path,
checkpoint_path=args.checkpoint_path)
# Push one empty image through to ensure Tensorflow is ready.
# There is typically a large wait on the first frame through.
model.predict(np.zeros(shape=[480, 640, 3]))
return model
def process(model, img):
steering_angle = model.predict(img)
if args.debug_print:
print(steering_angle)
return steering_angle
model = get_model()
print calc_rmse(lambda image_pred: model.predict(image_pred),
gen(args.bagfile))
return self.mean_angle[0]
else:
img = img1 - self.img0
img = rescale_intensity(img,
in_range=(-255, 255),
out_range=(0, 255))
img = np.array(img, dtype=np.uint8) # to replicate initial model
self.state.append(img)
self.img0 = img1
X = np.concatenate(self.state, axis=-1)
X = X[:, :, ::-1]
X = np.expand_dims(X, axis=0)
X = X.astype('float32')
X -= self.X_mean
X /= 255.0
return self.model.predict(X)[0][0]
if __name__ == "__main__":
parser = argparse.ArgumentParser(description='Model Runner for team rambo')
parser.add_argument('bagfile', type=str, help='Path to ROS bag')
args = parser.parse_args()
model = Model("tanel/final_model.hdf5", "tanel/X_train_mean.npy")
print calc_rmse(lambda image_pred: model.predict(image_pred),
gen(args.bagfile))