Can I find a way to reproduce the experiment result in paper?

[hsyoon94]

First, thank you for your great effort in your survey and code implementation : )

I am using it well for my research and I have a question. I tried to reproduce the result in your paper.

In Table 11 about ODI experiment, The average accuracy of ER with 1k memory for mini-imagenet-noise dataset is indicated as 19.4 however, in my experiment, it shows only about 10~11. Not only for this, others' performance show different in my experiment. Although I have tried to analyze the hyperparameters in your repo and applied them to mine, nothing special happened.

May I ask you for help reproducing the exact result in Table 11?

[hsyoon94]

I am reading your "Online Continual Learning in Image Classification: An empirical survey", Neurocomputing 2022. Thank you.

[Joe-Khawand]

Hello did you find a way ?