Higher test set accuracy?
cagbal opened this issue · 2 comments
Thank you for this nice repo and congrats. for your nice results. I have some questions about your repo.
As stated in #2, the test set accuracy is significantly higher than training accuracy during nearly first 100 epochs. Is it because of dropouts or any other reason?
Also, my second question is: are there any difference between eval accuracy during training mode and accuracy stated in --mode test? They are the same right?
My other question is: what is the difference between HCN[1] and HCN. Aren't they the same network with same configuration?
| HCN[1] | 32 | 256 | 86.5 | 91.1 |
|---|---|---|---|---|
| HCN | 32 | 256 | 84.2 | 89.2 |
Edit: I understand the first two questions, so I disabled them.
Sorry for the late reply. HCN[1] represents the results reported in the original paper, and HCN is our results. Because I just replicate it as a baseline for our work, we have not struggled to tune the parameters to meet the results reported in the original paper.