Leela Chess Zero
E771672
Leela Chess Zero is an open-source, neural-network-based chess engine inspired by AlphaZero that has become one of the strongest and most influential engines in computer chess.
Statements (49)
| Predicate | Object |
|---|---|
| instanceOf |
chess engine
ⓘ
computer chess engine ⓘ neural network chess engine ⓘ open-source software ⓘ |
| competesIn |
Computer chess tournaments
ⓘ
Top Chess Engine Championship NERFINISHED ⓘ |
| hasAbbreviation | Lc0 NERFINISHED ⓘ |
| hasAbbreviationMeaning | Lc0 stands for Leela Chess Zero NERFINISHED ⓘ |
| hasCommunity | Lc0 user and contributor community ⓘ |
| hasComponent |
client for distributed games
ⓘ
neural network weights ⓘ server for training ⓘ |
| hasDesignGoal | AlphaZero-like chess engine available to the public ⓘ |
| hasDocumentation | https://lczero.org/play/ ⓘ |
| hasForum | https://discord.gg/lczero ⓘ |
| hasNameOrigin | named after Futurama character Leela ⓘ |
| hasNotableFeature |
end-to-end neural network evaluation without handcrafted chess knowledge
ⓘ
policy and value network ⓘ support for multiple backends (CUDA, OpenCL, BLAS) ⓘ |
| hasSourceRepository | https://github.com/LeelaChessZero/lc0 ⓘ |
| hasStrength | top-tier engine strength comparable to strongest classical engines ⓘ |
| hasTrainingDataSource | self-play games generated by engine ⓘ |
| hasWebsite | https://lczero.org ⓘ |
| influenced | development of neural-network-based chess engines ⓘ |
| inspiredBy | AlphaZero NERFINISHED ⓘ |
| isMaintainedBy | community of volunteers ⓘ |
| isOpenSource | true ⓘ |
| isTrainedBy | distributed volunteer computing ⓘ |
| isWrittenFor | computer chess competitions ⓘ |
| license | GPL-3.0-or-later ⓘ |
| programmingLanguage |
C++
ⓘ
CUDA NERFINISHED ⓘ OpenCL NERFINISHED ⓘ |
| supportsHardware |
CPU
ⓘ
GPU ⓘ |
| supportsPlatform |
Linux
ⓘ
Windows ⓘ macOS ⓘ |
| supportsProtocol |
UCI
ⓘ
WinBoard NERFINISHED ⓘ |
| supportsTimeControl |
blitz chess
GENERATED
ⓘ
bullet chess GENERATED ⓘ classical chess GENERATED ⓘ rapid chess GENERATED ⓘ |
| trainingMethod | self-play reinforcement learning ⓘ |
| usesArchitecture | residual neural network ⓘ |
| usesEvaluation | deep neural network ⓘ |
| usesInputRepresentation | bitboard-like planes for neural network ⓘ |
| usesSearchAlgorithm | Monte Carlo tree search NERFINISHED ⓘ |
Referenced by (1)
Full triples — surface form annotated when it differs from this entity's canonical label.