A strong chess engine which uses a neural network to evaluate positions

Implementation of Medical Slice Transformer

TabFM (Tabular Foundation Model) is a pretrained tabular foundation model developed by Google Research for tabular data regression and classification.

Things you can do with the token embeddings of an LLM

CVChess is a state of the art deep learning framework with end to end chess state recognition from board state to FEN generation.

TinyStories-260K running locally on a stock Game Boy Color.

OCN — Open Chess Naming. Hierarchical names for chess openings, designed as a human-friendly companion to ECO.

A unified, torch-native training framework for MTP and block-parallel speculative decoding.

Sample source code games Atari 2600

A massively parallel, high-level programming language

Efficient Triton Kernels for LLM Training

PGN and chess game databases

Visa Vulnerability Agentic Harness

Python library for the Zulip API.

Color palettes for Python

GlobalBuildingAtlas: an open global and complete dataset of building polygons, heights and LoD1 3D models

This is the homepage of a new book entitled "Mathematical Foundations of Reinforcement Learning."

A massively multilingual corpus and pretrained model for IGT

Doom for computers with only 64 KB of RAM

A real-time, first-person raycaster demonstration for the SNK Neo Geo AES, written in C.

Browser-only chess practice app — Stockfish WASM, graded insights, spaced-repetition puzzles, bots, game review

A training implementation for Maia3 models

javascript multiplayer toolkit for building real-time games and applications

Automatic inference for composable models of heterogeneous data

DiRe accelerated by PyTorch, PyKeOps and cuVS

DiRe-RAPIDS + EVoC integration (Igor Rivin): GPU Phase-2 hybrid clustering, intrinsic dimension from the kNN graph, out-of-sample predict