[package]
edition = "2024"
rust-version = "1.97"
name = "hermes-tokenizer"
version = "1.8.102"
authors = ["izihawa"]
build = false
autolib = false
autobins = false
autoexamples = false
autotests = false
autobenches = false
description = "Stable-Rust byte-level BPE tokenization for Hermes"
readme = "README.md"
keywords = [
"tokenizer",
"bpe",
"llm",
"nlp",
]
categories = ["text-processing"]
license = "MIT"
repository = "https://github.com/SpaceFrontiers/hermes"
resolver = "2"
[lib]
name = "hermes_tokenizer"
path = "src/lib.rs"
[[test]]
name = "parity"
path = "tests/parity.rs"
[dependencies.aho-corasick]
version = "1.1"
[dependencies.anyhow]
version = "1.0"
[dependencies.icu_normalizer]
version = "2.2"
features = ["compiled_data"]
[dependencies.icu_properties]
version = "2.2"
features = ["compiled_data"]
[dependencies.memchr]
version = "2.8"
[dependencies.parking_lot]
version = "0.12"
[dependencies.rayon]
version = "1.11"
[dependencies.rustc-hash]
version = "2.1"
[dependencies.serde]
version = "1.0"
features = ["derive"]
[dependencies.serde_json]
version = "1.0"
features = ["float_roundtrip"]
[dev-dependencies.ahash]
version = "0.8"
[dev-dependencies.tokenizers]
version = "0.23"
features = ["fancy-regex"]
default-features = false
[target.'cfg(target_os = "linux")'.dependencies.libc]
version = "0.2"