The results of the paper report #7
Description
When using the code provided by the author to reproduce the paper, why is the result of reproduction two points lower than that reported in the paper? Is it some trick used in training?
{{ message }}
When using the code provided by the author to reproduce the paper, why is the result of reproduction two points lower than that reported in the paper? Is it some trick used in training?