[package]
edition = "2021"
name = "unillm-runtime"
version = "0.1.0"
build = false
autolib = false
autobins = false
autoexamples = false
autotests = false
autobenches = false
description = "Core inference runtime for UniLLM with 47 model architectures"
readme = "README.md"
license = "Apache-2.0"
repository = "https://github.com/cognisoc/unillm"
[features]
accelerate = [
"candle-core/accelerate",
"candle-nn/accelerate",
]
benchmark = ["llama-cpp-2"]
cuda = [
"candle-core/cuda",
"candle-nn/cuda",
]
default = []
gpu = [
"cuda",
"metal",
]
metal = [
"candle-core/metal",
"candle-nn/metal",
]
mkl = [
"candle-core/mkl",
"candle-nn/mkl",
]
performance = ["simd"]
simd = []
[lib]
name = "runtime"
path = "src/lib.rs"
[[bin]]
name = "benchmark_comparison"
path = "src/bin/benchmark_comparison.rs"
[[bin]]
name = "test_basic_inference"
path = "src/bin/test_basic_inference.rs"
[[bin]]
name = "test_ollama"
path = "src/bin/test_ollama.rs"
[[bin]]
name = "unillm"
path = "src/bin/unillm.rs"
[dependencies.aligned-vec]
version = "0.6"
[dependencies.anyhow]
version = "1.0"
[dependencies.async-trait]
version = "0.1"
[dependencies.axum]
version = "0.8"
features = [
"ws",
"multipart",
]
[dependencies.axum-extra]
version = "0.10"
features = ["typed-header"]
[dependencies.axum-test]
version = "15.0"
[dependencies.base64]
version = "0.22"
[dependencies.bincode]
version = "1.3"
[dependencies.blas]
version = "0.22"
[dependencies.bytemuck]
version = "1.18"
[dependencies.bytes]
version = "1.0"
[dependencies.candle-core]
version = "0.8"
[dependencies.candle-nn]
version = "0.8"
[dependencies.candle-transformers]
version = "0.8"
[dependencies.chrono]
version = "0.4"
features = ["serde"]
[dependencies.clap]
version = "4.0"
features = ["derive"]
[dependencies.dirs]
version = "5.0"
[dependencies.env_logger]
version = "0.11"
[dependencies.futures]
version = "0.3"
[dependencies.futures-util]
version = "0.3"
[dependencies.glob]
version = "0.3"
[dependencies.governor]
version = "0.6"
[dependencies.half]
version = "2.4"
[dependencies.image]
version = "0.25"
features = [
"png",
"jpeg",
"webp",
"tiff",
]
[dependencies.imageproc]
version = "0.25"
[dependencies.llama-cpp-2]
version = "0.1"
optional = true
[dependencies.lz4]
version = "1.24"
[dependencies.mime]
version = "0.3"
[dependencies.ndarray]
version = "0.16.1"
[dependencies.ndarray-linalg]
version = "0.16.0"
[dependencies.nonzero_ext]
version = "0.3"
[dependencies.num-traits]
version = "0.2"
[dependencies.once_cell]
version = "1.19"
[dependencies.prettytable-rs]
version = "0.10"
[dependencies.prometheus]
version = "0.13"
[dependencies.rand]
version = "0.8"
[dependencies.rayon]
version = "1.10"
[dependencies.reqwest]
version = "0.12"
features = [
"json",
"stream",
]
[dependencies.safetensors]
version = "0.6.2"
[dependencies.serde]
version = "1.0"
features = ["derive"]
[dependencies.serde_json]
version = "1.0"
[dependencies.sha2]
version = "0.10"
[dependencies.sysinfo]
version = "0.32"
[dependencies.tempfile]
version = "3.8"
[dependencies.thiserror]
version = "2.0"
[dependencies.tokenizers]
version = "0.20"
[dependencies.tokio]
version = "1.0"
features = ["full"]
[dependencies.tokio-tungstenite]
version = "0.24"
[dependencies.tower]
version = "0.5"
[dependencies.tower-http]
version = "0.6"
features = [
"cors",
"fs",
"trace",
"timeout",
"compression-gzip",
]
[dependencies.tracing]
version = "0.1"
[dependencies.tracing-subscriber]
version = "0.3"
features = [
"env-filter",
"json",
]
[dependencies.uuid]
version = "1.0"
features = ["v4"]
[dependencies.wide]
version = "0.7"