search

Sujit-O / pykg2vec

Python library for knowledge graph embedding and representation learning.
MIT License
602 stars 109 forks source link

Optimal parameter selection #214

Open HL-97 opened 3 years ago

HL-97 commented 3 years ago

In the process of tuning parameters, I found that after training the same number of epochs (only tested 10 times), the golden setting returned by Bayesian optimization is not the one with the smallest loss or the largest hit value. What is the cause of this? Thanks for your answer！

HL-97 commented 3 years ago

So should I choose the golden setting returned by Bayesian optimization, or manually select the set of parameters with the lowest loss and the largest hits value by myself ?