Akarsh Kumar's repositories
evocraft-2022
GECCO Open-Endedness Challenge
transformer-memory
A transformer with a long-term memory mechanism
akarshkumar.com
my personal website
akarshkumar0101
My Github Homepage!
auto-pic
Automated Picbreeder!
basketball-rl
Basketball Play Simulation with Reinforcement Learning!
cluster-run
Scripts for launching many jobs on many remote servers.
dotfiles
My personal config files! Has files for .bashrc, .zshrc, .vimrc, .gitconfig, .tmux.conf, etc.
ee-422-rec6
Pair Programming Recitation
evolved-neuroevolution
Making neuroevolution more end-to-end evolved by evolving the strategy of mutating and crossing over.
exploration-generalization
Exploration generalizes (hopefully better!)
hardware-final-project
Hardware final project
meta-weight-sharing
Finding the optimal weight-sharing technique in NNs using meta-learning!
minecraft_server
Our Minecraft Server! Akarsh, Louis, Corey, Sarkis, Sahran
mlp-shaker
Code for MLP-Shaker, an extension on MLP-Mixer
PPO-PyTorch
Minimal implementation of clipped objective Proximal Policy Optimization (PPO) in PyTorch
procgen
Procgen Benchmark: Procedurally-Generated Game-Like Gym-Environments
pytorch3d
PyTorch3D is FAIR's library of reusable components for deep learning with 3D data
qd-acl
Quality Diversity for Automatic Curriculum Learning!
retnet-test
The simplest, fastest repository for training/finetuning medium-sized GPTs.
sparse-high-attention
Sparse Higher-Order Attention Mechanism
synthetic-mdps
Synthetic MDPs
timm-mlp-shaker
PyTorch image models, scripts, pretrained weights -- ResNet, ResNeXT, EfficientNet, EfficientNetV2, NFNet, Vision Transformer, MixNet, MobileNet-V3/V2, RegNet, DPN, CSPNet, and more
true-neutral-cookiecutter
A minimalist cookiecutter for research and data science projects
two-man-sgd
Finding the (approximate) intersection between two point-gradient rays of the parent NNs in weight space as a crossover mechanism. Can be used as a neuroevolution crossover or be adapted for an optimization algorithm.