Model weights (trained on ImageNet Subset)

Contains weights for the experiments performed with two denoising autoencoders with varying depth:

• Autoencoder weights
• SimCLR (our variant) weights
• G-SimCLR weights

Model weights (trained on CIFAR10)

Contains:

• Autoencoder weights
• Vanilla SimCLR (our variant) weights
• G-SimCLR weights

AutoencoderData

The .npy files are strictly for the purposes of fast-tracking our experiments with the image net subset.
X_new.npy is the ordered training dataset (Imagenet Subset). The ordering takes place as per the deep autoencoder latent space clusters.
X_new_shallow.npy is the ordered training dataset (Imagenet Subset). The ordering takes place as per the shallow autoencoder latent space clusters.