a-mch

Thanks for suggestion, I have tried extracting, ```python @tf.function def compute_gradient(model_): with tf.GradientTape() as g: loss_flow = model_.training_loss() return g.gradient(loss_flow, model_.trainable_variables) ``` and calling this function in the loop but...

Hi, thanks for looking into it. I tried computing derivatives of a sq exp covariance function written with TensorFlow functions, like so, ``` import tensorflow as tf import time def...