TensorRT result is different from Tensorflow results

Hi, I am running the Lanenet (GitHub - MaybeShewill-CV/lanenet-lane-detection: Unofficial implemention of lanenet model for real time lane detection using deep neural network model https://maybeshewill-cv.github.io/lanenet-lane-detection/). I first train it on my desktop with tensorflow and python. Then I convert the tensorflow checkpoint to uff and run inference with tensorrt on Jetson Xavier. However, I found the results from tensorflow python and c++ tensort rt are slightly different. Please see the two picture below. Does anyone know why?

rear_0012-real
Tensorflow python result

rear_0012_tensorrt
TensorRT c++ result

wroking on heath data set from vitalefy, i used to get the same error, converting input to np.float32 solved my issues

input_img = input_img.astype(np.float32)