rlx-dinov2 0.2.4

DINOv2 ViT encoder for RLX
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
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

// 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/>.

// RLX CLI
use crate::{DinoV2Output, DinoV2Runner, DinoV2Variant};
use anyhow::{Context, Result, anyhow, bail};
use rlx_cli::{parse_standard_device, req};
use std::path::PathBuf;
pub fn run(args: &[String]) -> Result<()> {
    let mut weights: Option<PathBuf> = None;
    let mut device = "cpu".to_string();
    let mut variant_str = "base".to_string();
    let mut img_size = 518usize;
    let mut batch = 1usize;
    let mut dry = false;
    let mut i = 0;
    while i < args.len() {
        match args[i].as_str() {
            "--weights" => {
                weights = Some(req(args, &mut i)?.into());
            }
            "--device" => device = req(args, &mut i)?,
            "--variant" => variant_str = req(args, &mut i)?,
            "--img-size" => {
                img_size = req(args, &mut i)?.parse().context("--img-size: usize")?;
            }
            "--batch" => batch = req(args, &mut i)?.parse().context("--batch: usize")?,
            "--dry" => {
                dry = true;
                i += 1;
            }
            "--help" | "-h" => {
                eprintln!("rlx-dinov2 — see README for flags");
                return Ok(());
            }
            other => bail!("unknown flag: {other}"),
        }
    }
    let weights = weights.ok_or_else(|| anyhow!("--weights is required"))?;
    let device = parse_standard_device("dinov2", &device)?;
    let variant = match variant_str.as_str() {
        "small" | "vit-s" => DinoV2Variant::Small,
        "base" | "vit-b" => DinoV2Variant::Base,
        "large" | "vit-l" => DinoV2Variant::Large,
        other => bail!("--variant: expected small|base|large (got {other})"),
    };

    eprintln!(
        "[rlx-dinov2] dinov2: weights={weights:?} device={device:?} variant={variant:?} img_size={img_size} batch={batch}"
    );
    let mut runner = DinoV2Runner::builder()
        .weights(&weights)
        .device(device)
        .variant(variant)
        .img_size(img_size)
        .batch(batch)
        .build()?;
    eprintln!(
        "[rlx-dinov2] compiled — hidden={} layers={} num_classes={}",
        runner.config().hidden_size,
        runner.config().num_hidden_layers,
        runner.config().num_classes
    );

    if dry {
        eprintln!("[rlx-dinov2] --dry set; skipping forward pass");
        return Ok(());
    }

    // Synthetic image — replace with image::open(...).to_rgb8().as_raw().
    let (h_in, w_in) = (img_size, img_size);
    let mut rgb = vec![0u8; h_in * w_in * 3];
    for y in 0..h_in {
        for x in 0..w_in {
            let base = (y * w_in + x) * 3;
            rgb[base] = (x * 255 / w_in) as u8;
            rgb[base + 1] = (y * 255 / h_in) as u8;
            rgb[base + 2] = ((x + y) * 127 / (h_in + w_in)) as u8;
        }
    }

    let t0 = std::time::Instant::now();
    let out = runner.predict_image(&rgb, h_in, w_in)?;
    let dt = t0.elapsed();
    match out {
        DinoV2Output::Logits {
            per_batch,
            num_classes,
        } => {
            eprintln!(
                "[rlx-dinov2] dinov2 logits in {dt:?} — batch={} classes={}",
                per_batch.len(),
                num_classes
            );
            for (b, logits) in per_batch.iter().enumerate() {
                let (top1, top1_val) = logits
                    .iter()
                    .enumerate()
                    .max_by(|(_, a), (_, b)| a.total_cmp(b))
                    .unwrap();
                eprintln!("  batch[{b}] top1={top1} logit={top1_val:.3}");
            }
        }
        DinoV2Output::Tokens {
            per_batch,
            seq,
            hidden,
        } => {
            eprintln!(
                "[rlx-dinov2] dinov2 tokens in {dt:?} — batch={} seq={seq} hidden={hidden}",
                per_batch.len()
            );
            // CLS token (index 0) summary: ||cls||₂
            for (b, toks) in per_batch.iter().enumerate() {
                let cls = &toks[..hidden];
                let norm: f32 = cls.iter().map(|x| x * x).sum::<f32>().sqrt();
                eprintln!("  batch[{b}] ||cls||₂ = {norm:.3}");
            }
        }
    }
    Ok(())
}