laya 0.1.1

Rust inference for the Laya non-autoregressive typed-decision model (ModernBERT-large + RL decision head)
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46

[package]
name = "laya"
version = "0.1.1"
edition = "2021"
rust-version = "1.75"
authors = ["Enzo Lombardi <enzinol@gmail.com>"]
description = "Rust inference for the Laya non-autoregressive typed-decision model (ModernBERT-large + RL decision head)"
repository = "https://github.com/aovestdipaperino/laya-rust"
license = "Apache-2.0"
readme = "README.md"
keywords = ["nlp", "classification", "candle", "modernbert", "inference"]
categories = ["science"]
# Keep the package lean: the demo screenshots and the checkpoint stay in git / on the Hub.
exclude = ["models/*", "examples/_*_tmp.rs"]

[dependencies]
candle-core = "0.9"
candle-nn = "0.9"
candle-transformers = "0.9"
tokenizers = "0.21"
serde = { version = "1", features = ["derive"] }
serde_json = { version = "1", features = ["preserve_order"] }
anyhow = "1"
clap = { version = "4", features = ["derive"] }
axum = { version = "0.8", optional = true }
tokio = { version = "1", features = ["rt-multi-thread", "macros", "signal", "net"], optional = true }

[features]
default = []
# The web console binary. Opt in: it pulls axum and tokio.
serve = ["axum", "tokio"]
metal = ["candle-core/metal", "candle-nn/metal", "candle-transformers/metal"]
cuda = ["candle-core/cuda", "candle-nn/cuda", "candle-transformers/cuda"]

[[bin]]
name = "laya"
path = "src/bin/laya.rs"

[[bin]]
name = "laya-serve"
path = "src/bin/laya-serve.rs"
required-features = ["serve"]

[profile.release]
opt-level = 3
lto = "thin"