Specifying separate tf.summary file_writers for training and validation

Question

I would like to visualize some scalars such as Train loss, val loss (among others) as my training progresses. I am using tf.Keras with tf2.3. However I am unable to specify