Song Bian

I apologize for my delayed response since I am busy recently. It seems that the latest version of PyTorch does not support our code. Could you use pytorch 1.0? because...

@jinglinpeng n-gram frequency

I think the reason is that plot_correlation does not deal with nan/null value. Thus, we have to consider nan/null value in the plot_correlation function? We have to deal with nan/null...

I wonder whether Peter has implemented drop_null function? Or I just use df.dropna() first, then I replace it with drop_null function after Peter implements it? @dovahcrow @jinglinpeng

Thanks for your reply! I modify the code as follows: ``` python import numpy as np import tensorflow as tf from tensorflow.keras.datasets import cifar10 import epl import os def conv_bn_relu(inputs,...

Thanks for your reply! I have run the main function in train.py - benchmark_task_val(), but the *max validation* performance (there is no test set performance) overall training iterations enzymes are...

Thanks for your reply! I have modified the hyper-parameters. And I have run the main function in train.py - benchmark_task_val() for sevaral times. The result of DD and enzymes have...