Entropy Based Sampling and Parallel CoT Decoding

DIAMOND (DIffusion As a Model Of eNvironment Dreams) is a reinforcement learning agent trained in a diffusion world model. NeurIPS 2024 Spotlight.

Repo to submit jobs to the AMD cluster

Open EDR public repository

Useful scripts to exploit Hack The Box retired machines/challenges

Resizable BAR for (almost) any UEFI system

The Network Execution Tool

System 2 Reasoning Link Collection

Detect file content types with deep learning

Felafax is building AI infra for non-NVIDIA GPUs

A playbook for systematically maximizing the performance of deep learning models.

High performance AI inference stack. Built for production. @ziglang / @openxla / MLIR / @bazelbuild

Library for Jacobian descent with PyTorch. It enables optimization of neural networks with multiple losses (e.g. multi-task learning).

Neo AI integrates into the Linux terminal, capable of executing system commands and providing helpful information.

Minimal Prolog implemented in Rust

The NetHack Learning Environment

The AdEMAMix Optimizer: Better, Faster, Older.

The Roguelike Toolkit (RLTK), implemented for Rust.

2D Platformer Educational Game for Teaching Game Hacking - C /cocos2d-x

Squalr Memory Editor - Game Hacking Tool Written in C#

[ACL'24] Selective Reflection-Tuning: Student-Selected Data Recycling for LLM Instruction-Tuning

24 channel, 100Msps logic analyzer hardware and software

Simplifying reinforcement learning for complex game environments

94% on CIFAR-10 in 2.6 seconds 💨 96% in 27 seconds

Beyond Language Models: Byte Models are Digital World Simulators

Autonomous Agents (LLMs) research papers. Updated Daily.

Demonstrations of Loss of Plasticity and Implementation of Continual Backpropagation