Giters

frank-xwang / RIDE-LongTailRecognition

[ICLR 2021 Spotlight] Code release for "Long-tailed Recognition by Routing Diverse Distribution-Aware Experts."

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Request for more results

Harry-hash opened this issue · comments

commented

It seems that currently only top-1 accuracy is reported after training on imagenet-LT, and I also see the script to compute top-k accuracy. I am wondering how can I get the performance on many-shot, medium-shot and few-shot?

commented

Hi, we have provided the script for testing the pre-trained checkpoints, please follow "Test" section of the instruction in README. The results reported during the training process is only for the sake of debugging. You can also modify it to report top-5 by calling top_k_acc function in RIDE-LongTailRecognition/model/metric.py and set k as 5. Feel free to re-oepn this issue and let us know if you have more problems.