rlx-onnx 0.2.4

ONNX inference for RLX — native compile by default, optional ORT fallback
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
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151

// RLX — versatile ML compiler + runtime.
// Copyright (C) 2026 Eugene Hauptmann, Nataliya Kosmyna.
//
// This program is free software: you can redistribute it and/or modify
// it under the terms of the GNU General Public License as published by
// the Free Software Foundation, version 3.
//
// This program is distributed in the hope that it will be useful,
// but WITHOUT ANY WARRANTY; without even the implied warranty of
// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
// GNU General Public License for more details.
//
// You should have received a copy of the GNU General Public License
// along with this program. If not, see <https://www.gnu.org/licenses/>.

//! CLI: load an ONNX file and run one inference (zero-filled inputs by default).

use std::env;
use std::path::PathBuf;
use std::time::Instant;

use anyhow::{Context, Result, bail};
use rlx_onnx::{OnnxCompileLevel, OnnxExecBackend, OnnxModel};
use rlx_runtime::{Device, parse_device};

fn usage() -> &'static str {
    "usage: rlx-onnx-run <model.onnx> [--device cpu|cuda|metal|...] [--level 0-3] [--exec native|ort] [--list-io] [--seq-len N] [--warmup N] [--iters N]"
}

fn main() -> Result<()> {
    let mut args = env::args().skip(1);
    let model = args
        .next()
        .map(PathBuf::from)
        .filter(|p| p.extension().is_some_and(|e| e == "onnx"))
        .context(usage())?;

    let mut device = Device::Cpu;
    let mut list_io = false;
    let mut warmup = 1usize;
    let mut iters = 1usize;
    let mut seq_len = 128usize;
    let mut level = OnnxCompileLevel::Level3;
    let mut backend = OnnxExecBackend::Native;

    while let Some(a) = args.next() {
        match a.as_str() {
            "--device" => {
                let s = args.next().context("--device requires a value")?;
                device = parse_device(&s)?;
            }
            "--level" => {
                let n: u8 = args
                    .next()
                    .context("--level 0-3")?
                    .parse()
                    .context("--level 0-3")?;
                level = OnnxCompileLevel::from_u8(n);
            }
            "--exec" => {
                let s = args.next().context("--exec native|ort")?;
                backend = match s.as_str() {
                    "native" | "rlx" => OnnxExecBackend::Native,
                    "ort" => {
                        #[cfg(feature = "ort")]
                        {
                            OnnxExecBackend::Ort
                        }
                        #[cfg(not(feature = "ort"))]
                        {
                            bail!(
                                "rlx-onnx-run: rebuild with feature `ort-fallback` for --exec ort"
                            )
                        }
                    }
                    other => bail!("unknown --exec {other} (use native or ort)"),
                };
            }
            "--list-io" => list_io = true,
            "--warmup" => {
                warmup = args
                    .next()
                    .context("--warmup N")?
                    .parse()
                    .context("--warmup N")?;
            }
            "--iters" => {
                iters = args
                    .next()
                    .context("--iters N")?
                    .parse()
                    .context("--iters N")?;
            }
            "--seq-len" => {
                seq_len = args
                    .next()
                    .context("--seq-len N")?
                    .parse()
                    .context("--seq-len N")?;
            }
            "--help" | "-h" => {
                println!("{}", usage());
                return Ok(());
            }
            other => bail!("unknown argument: {other}"),
        }
    }

    let mut model = OnnxModel::load_with(&model, device, backend, level, seq_len)?;
    if list_io {
        model.print_io();
        return Ok(());
    }

    let inputs = model.zero_inputs_sized(seq_len as i64)?;
    for _ in 0..warmup {
        let _ = model.run(&inputs)?;
    }

    let t0 = Instant::now();
    let mut last = None;
    for _ in 0..iters {
        last = Some(model.run(&inputs)?);
    }
    let elapsed = t0.elapsed();

    let outs = last.context("no iterations")?;
    let backend_s = format!("{:?}", model.backend);
    let ep = model
        .ort_ep
        .as_deref()
        .map(|e| format!(", ort_ep={e}"))
        .unwrap_or_default();
    println!(
        "ok: {} output(s), {} iter(s), {:.2} ms total ({:.2} ms/iter), backend={backend_s}{ep}, level={:?}",
        outs.len(),
        iters,
        elapsed.as_secs_f64() * 1000.0,
        elapsed.as_secs_f64() * 1000.0 / iters as f64,
        model.compile_level,
    );
    for (desc, tensor) in model.outputs.iter().zip(outs.iter()) {
        let n = match tensor {
            rlx_onnx::OnnxTensor::F32(v) => v.len(),
            rlx_onnx::OnnxTensor::I64(v) => v.len(),
            rlx_onnx::OnnxTensor::I32(v) => v.len(),
        };
        println!("  output {}: {} elements", desc.name, n);
    }
    Ok(())
}