gboduljak

vae

Implementations of various VAEs.

mlx

MLX: An array framework for Apple silicon

mlx-examples

Examples in the MLX framework

mlx-data

Efficient framework-agnostic data loading

group-equivariant-cnns-vs-spatial-transformers

This study aims to show that group equivariant CNNs outperform spatial transformers, on tasks which demand rotation invariance, by providing theoretical background and experimental performance comparison with detailed analysis.

lagan

Lightweight attentional GAN for unpaired image-to-image-translation.

pytorch-CycleGAN-and-pix2pix

Image-to-Image Translation in PyTorch

contrastive-unpaired-translation

Contrastive unpaired image-to-image translation, faster and lighter training than cyclegan (ECCV 2020, in PyTorch)

UGATIT-pytorch

Official PyTorch implementation of U-GAT-IT: Unsupervised Generative Attentional Networks with Adaptive Layer-Instance Normalization for Image-to-Image Translation

query-selected-attention

Official implementation for "QS-Attn: Query-Selected Attention for Contrastive Learning in I2I Translation" (CVPR 2022)

NICE-GAN-pytorch

Official PyTorch implementation of NICE-GAN: Reusing Discriminators for Encoding: Towards Unsupervised Image-to-Image Translation

stargan-v2

StarGAN v2 - Official PyTorch Implementation (CVPR 2020)

pixel-da

pixel da

CyCADA

A PyTorch implementation of CyCADA

DRIT

Learning diverse image-to-image translation from unpaired data

SoloGAN

Official implementation of SoloGAN

TiTok

Code for the paper TIToK: A solution for bi-imbalanced unsupervised domain adaptation

graph-attention-networks

An implementation of Graph Attention Networks (GAT) paper accompanied with the analysis of attention scores.

bsc-thesis

The thesis is "On universality of fully-connected neural networks". This thesis aims to present key results in the approximation theory of artificial neural networks assuming only undergraduate mathematics. This research field studies necessary and sufficient conditions under which neural networks can approximate certain functions.

MUNIT

Multimodal Unsupervised Image-to-Image Translation

stanford-compilers-coursework

This is a repository of coursework project for the Stanford Compilers MOOC course. The result is a fully-working compiler for the COOL Programming Language.

seam-carving

This is my implementation of the simplified version of the seam carving - an algorithm developed by S.Avidan and A.Shamir. Seam carving is an algorithm for 'content-aware' image resizing. The main idea is to resize an image by removing only the least noticeable pixels. I have also made a small web app used to generate interesting results.

mnist-from-scratch

MNIST done with paper, pencil and Numpy

LeafGAN

mcc

mcc is a toy compiler compiling a (relatively) large subset of C language to LLVM IR.

prettyprinter

A modern, extensible and well-documented prettyprinter.

uscheme

μscheme (micro Scheme) is a toy interpreter for the Scheme programming language, implementing a relatively (large) subset of R5RS language specification.

functional-pearls

Implementations of functional pearls (https://wiki.haskell.org/Research_papers/Functional_pearls) I read.

dft

A quick and dirty implementation of Discrete Fourier transform (DFT) in Haskell. Tested on Polynomial Multiplication.

calc

A console based calculator written in Haskell.