Skip to content

NikkeTryHard/hydra

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

765 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Hydra

Open-source Riichi Mahjong AI. The goal is to build an AI that rivals LuckyJ (Tencent, 10.68 stable dan on Tenhou) with open weights.

Goal

Train a mahjong AI that:

  • Surpasses Mortal (~7-dan) and approaches LuckyJ-level play (10+ dan) in head-to-head evaluation
  • Releases weights under a permissive license
  • Adds opponent modeling and inference-time search — the two capabilities that separate LuckyJ from all other mahjong AIs

Architecture

Hydra's strongest current doctrine is:

  1. research/design/HYDRA_FINAL.md — target architecture north star
  2. research/design/HYDRA_RECONCILIATION.md — current execution doctrine
  3. docs/GAME_ENGINE.md — current runtime reality

Everything else should be read as implementation reference, rationale, reserve shelf, or archive material unless one of the three files above explicitly promotes it.

If you are deciding what to build next, read these in order:

research/design/HYDRA_SPEC.md is historical context only.

Research

File What's In It
HYDRA_FINAL.md Current target architecture SSOT
HYDRA_RECONCILIATION.md Current execution doctrine and active/reserve split
HYDRA_ARCHIVE.md Archived / reserve-only planning surfaces
HYDRA_SPEC.md Historical architecture spec only
MORTAL_ANALYSIS.md Mortal's architecture, training details, confirmed weaknesses
OPPONENT_MODELING.md Opponent-modeling rationale; includes both active ideas and reserve/future extensions
INFRASTRUCTURE.md Rust stack, data pipeline, training infra, hardware, deployment
SEEDING.md RNG hierarchy, reproducibility, evaluation seed bank
CHECKPOINTING.md Checkpoint format, save protocol, retention policy
ECOSYSTEM.md Useful repos, tooling, and framework references
REWARD_DESIGN.md Reward design and RVR notes
COMMUNITY_INSIGHTS.md Community observations and external signals
REFERENCES.md Citation index
TESTING.md Testing strategy, correctness verification, property-based tests
RUST_STACK.md 100% Rust decision and framework notes

Documentation Ownership (SSOT)

  • research/design/HYDRA_FINAL.md: target architecture SSOT
  • research/design/HYDRA_RECONCILIATION.md: current repo-wide decision memo and best-next-action guide
  • docs/GAME_ENGINE.md: current game-engine/runtime baseline
  • research/design/OPPONENT_MODELING.md: detailed opponent-modeling rationale
  • research/design/HYDRA_ARCHIVE.md: archive / reserve-only doctrine
  • research/design/HYDRA_SPEC.md: historical architecture summary only; do not use it as the current implementation SSOT

Status

Active implementation. hydra-core is already built out as a real baseline engine/encoder crate, and hydra-train contains a substantial training/model scaffold with partial advanced integration. The immediate project need is reconciliation plus closure of the highest-leverage training/search supervision loops, not a restart from scratch.

License

  • hydra-core (encoder, training pipeline): BSL 1.1 -- free for non-commercial use, converts to Apache-2.0 on 2031-03-02
  • hydra-engine (game rules): Apache-2.0 (vendored from riichienv-core)

About

No description, website, or topics provided.

Resources

License

Stars

Watchers

Forks

Packages

 
 
 

Contributors