Okyumi/Okyumi.github.io
Yumi Omori's research and personal writing.
Doing research @nyuad. Lifelong sidequester :)
Repositories
Okyumi/iwanna-gym
RL environment with exact I Wanna fangame physics: 5.8M steps/s C core, PufferLib-ready, goal-conditioned (HER/GCRL) support, 10 levels
Okyumi/endwalker_2026
Okyumi/builderbench-stablecrl-dcc
StableCRL BuilderBench with permutation-aware DCC continual reinforcement learning.
Okyumi/sgcrl
Okyumi/OverTrack
Overland travel route planner — from Tokyo to Antarctica, no flights required.
Okyumi/builderbench
Okyumi/CCRL_Research_Journals
Okyumi/scaling-crl
Okyumi/DeepScientist
Now, Stronger AI Pushes Frontiers, Stronger Our Shared Future.
Okyumi/demystifying-single-goal-exploration-code
Codebase of the paper "Demystifying the Mechanisms Behind Emergent Exploration in Goal-conditioned RL"
Okyumi/cgf
Official website for the China-Gulf Forum (CGF) — a student-led platform fostering China-Gulf dialogue since 2019 | china-gulf-forum.org
Okyumi/overtrack-claw
OpenClaw plugin — plan international overland routes by train, bus, and ferry. 191 cities · 506 connections · 103 operators · 22 countries · Dijkstra pathfinding
Okyumi/CLI-Anything
CLI-Anything: Making ALL Software Agent-Native
Okyumi/OverTrack-iOS
OverTrack — Overland international travel route planner. React Native (Expo) iOS app.
Okyumi/open-swe
An Open-Source Asynchronous Coding Agent
Okyumi/crl-chess-go
Contrastive RL for Board Games: experiments applying InfoNCE-based GCRL to Chess, Go, and Connect-4
Okyumi/Gymnasium
An API standard for single-agent reinforcement learning environments, with popular reference environments and related utilities (formerly Gym)
Okyumi/cleanrl
High-quality single file implementation of Deep Reinforcement Learning algorithms with research-friendly features (PPO, DQN, C51, DDPG, TD3, SAC, PPG)
Okyumi/AutoResearchClaw
Fully autonomous & self-evolving research from idea to paper. Chat an Idea. Get a Paper. 🦞
Okyumi/featured-blogs
Featured blog posts from omoriyumi.com
Okyumi/Okyumi
Okyumi/pdf-tool
A fast, private, client-side PDF utility for splitting, merging, and compressing PDFs. No uploads, no servers — everything runs in your browser.
Okyumi/endwalker
ENDWALKER — A parametric design sculpture exploring entropy, solar geometry, and generative form. Interactive project website by Yumi Omori, NYU Abu Dhabi.
Okyumi/cognitive-neuroscience-notes
Structured notes from PSYCH-UH 2412 Cognitive Neuroscience at NYU Abu Dhabi. Covers perception, attention, working memory, long-term memory, and consciousness.
Okyumi/CKA-RL
Continual Knowledge Adaptation for Reinforcement Learning
Okyumi/Machine-Lab
Okyumi/CCRL
Okyumi/annotated_deep_learning_paper_implementations
🧑🏫 60+ Implementations/tutorials of deep learning papers with side-by-side notes 📝; including transformers (original, xl, switch, feedback, vit, ...), optimizers (adam, adabelief, sophia, ...), gans(cyclegan, stylegan2, ...), 🎮 reinforcement learning (ppo, dqn), capsnet, distillation, ... 🧠
Okyumi/CORL
High-quality single-file implementations of SOTA Offline and Offline-to-Online RL algorithms: AWAC, BC, CQL, DT, EDAC, IQL, SAC-N, TD3+BC, LB-SAC, SPOT, Cal-QL, ReBRAC