Official implementation for "Pure Noise to the Rescue of Insufficient Data: Improving Imbalanced Classification by Training on Random Noise Images" https://arxiv.org/abs/2112.08810
I have read the paper on Pure Noise to the Rescue of Insufficient Data and found it very simple and exciting. However, I have the following questions:
When will the code be available?
Are the experimental comparisons fair? The authors use different network architectures on CIFAR-10/-100 and ImageNet, compared to the commonly-used setting.
Have the authors tried to experiment with ResNet-32 on CIFAR-10/-100๏ผ I tried reproducing the code but didn't get better results than MiSLAS.