G-SSL

Gated Self-Supervised Learning

[Arxiv]

This is an official implementation for "Mixture of Self-Supervised Learning". This implementations only provide Tiny-Imagenet classification, imbalanced classification and adversarial perturbations.

This is the diagram of the transformation from our proposed method.

This is the architecture of the model from our proposed method.

Dataset

• Tiny-Imagenet link
• CIFAR link

Requirements

• Python 3.10
• Pytorch 2.x
• Pyyaml
• Tensorboard
• Matplotlib