search

Yunfan-Li / Contrastive-Clustering

Code for the paper "Contrastive Clustering" (AAAI 2021)
MIT License
297 stars 91 forks source link

ImageNet 10 #20

Open YIBO-0805 opened 2 years ago

YIBO-0805 commented 2 years ago

Hi，

I did not find a description about the distribution of the Imagenet-10 and the data-augment method you used in your paper and code. And I get a bad accuracy on this dataset.

I would like to know if you could describe in detail or if I get something wrong？

Yunfan-Li commented 2 years ago

Hi, ImageNet-10 is a subset of ImageNet with 10 class indices provided in the dataset folder. The data augmentation for ImageNet-10 could be found at https://github.com/Yunfan-Li/Contrastive-Clustering/blob/dfda491ce22541265228e119d7e708fe752d4ad7/modules/transform.py#L21.

YIBO-0805 commented 2 years ago

Thank you for your reply. Following the dataset folder, I got 14,216 images. I would like to know how you screen out 13,000 images and how many images are in each category.

Yunfan-Li commented 2 years ago

In our experiment, the training set of ImageNet is used with 1300 images in each class.