I ran the code without any modification and the reproduced result was 0.4 lower than the reported 76.04 (0.2 lower if using provided pretrained teacher model). Could you help me figure whether the result was acceptable? What can I do to increase the reproduced result? Thanks.
Hi,
I ran the code without any modification and the reproduced result was 0.4 lower than the reported 76.04 (0.2 lower if using provided pretrained teacher model). Could you help me figure whether the result was acceptable? What can I do to increase the reproduced result? Thanks.