메뉴
Use when working on XFeat pre/post-processing, keypoints, descriptors, or matching — tensor shapes, coordinate spaces, normalization, NMS/top-K, and the exact algorithm steps implemented in vrt-xfeat.
기본적으로 소스를 먼저 확인하는 Prompt가 선택됩니다. 직접 명령으로 전환하거나 로컬 사본을 다운로드할 수도 있습니다.
설치 여부를 결정하기 전에 SKILL.md와 SkillsMP에 표시된 보조 파일을 읽어 보세요.
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
직접 명령은 검토 Prompt를 거치지 않습니다. 실행하기 전에 소스를 확인하세요.
npx skills add https://github.com/kornia/vision-rt --skill model-tensor-semantics명령은 한 줄로 유지됩니다. 복사하기 전에 가로로 스크롤해 전체 내용을 확인하세요.
로컬 사본을 원하시나요? SkillsMP에서 현재 제공할 수 있는 파일을 다운로드하세요.
SOC 직업 분류 기준
SKILL.md 표시 중
| name | model-tensor-semantics |
| description | Use when working on XFeat pre/post-processing, keypoints, descriptors, or matching — tensor shapes, coordinate spaces, normalization, NMS/top-K, and the exact algorithm steps implemented in vrt-xfeat. |
XFeat is a plain Image<u8,3> → XFeatResult type (no pipeline/operator
framework). Device data crosses boundaries as kornia types + TRT views:
Image<u8,3> (device), any resolution. XFeat owns a
kornia Preprocessor::stretch and resizes each frame to its own floor-of-32
dims mh,mw = (H/32)*32, (W/32)*32 into a reused Tensor<f32,4> [1,3,mh,mw]
(reallocated only when the size changes).ModelSession::run(&Tensor<f32,4>) -> TRTensorMap; outputs
read by name via TRTensorMap::get("descriptors"|"heatmap"|"reliability") →
OutputView::f32_ptr() (dtype-checked device pointer, valid until the next
run). See crates/vrt-xfeat/src/model.rs.XFeatResult with
XFeat::alloc_result(), then xfeat.submit(&img, &mut result) (async, no sync)
→ stream.sync() → read. run() = alloc + submit + sync convenience. Hold
several results to keep multiple frames outstanding under one sync.XFeatResult (device kpts/descs/scores, capacity ) —
reads the pinned scalar (post-sync), applies the
(returns original pixels), is a plain D2H.top_kcount()kpts_to_hostscale (rw,rh)scores_to_hostThis mirrors upstream XFeat exactly (preprocess_tensor):
XFeatResult.kpts).XFeatResult.scale = (rw, rh) = (W/mw, H/mh) maps model→original. kpts_to_host
applies it, so host keypoints are in original pixels (upstream's
mkpts * [rw, rh]). The resize is anisotropic but sub-32px, so aspect is
effectively preserved without any padding — no letterbox, no pad offset.
Descriptor matching uses descriptors only, so device kpts staying in model
space doesn't affect it.
Preprocessor::stretch)[1,3,mh,mw], values /255 → [0,1] (no mean/std),
anisotropic resize to floor-of-32 (matches XFeat's F.interpolate).XFeatParams is
just { top_k, threshold } — the input size is per-frame, not configured.| Tensor | Shape | Meaning |
|---|---|---|
descriptors | (1, 64, H/8, W/8) | dense 64-D feature map |
heatmap | (1, 1, H, W) | keypoint confidence |
reliability | (1, 1, H, W) | per-pixel reliability weight |
64 is XFeat's own width. The matcher is separate and compiles for others — see below.
Engine MUST expose exactly those three output names (model.rs errors with
MissingOutput otherwise).
crates/vrt-xfeat/src/postprocess.rs, NVRTC-JIT kernels)xfeat_score_nms — 5×5 local-max NMS; score = heatmap×reliability, zeroed
below params.threshold or if any neighbour is greater.xfeat_topk_histogram bins survivor scores,
xfeat_topk_cutoff finds the score cutoff for ~K survivors, xfeat_topk_select
atomically gathers survivors ≥ cutoff, capped at params.top_k. Approximate
only at the boundary bucket (1024 bins). Output is atomic-append order,
NOT score-sorted.xfeat_sample_descs — bilinear-sample 64-D descriptors at kpt/8 positions,
align_corners=False (matches PyTorch grid_sample).xfeat_l2_norm — in-place L2-normalize each descriptor row.Async contract: submit() enqueues preproc → backbone → NMS → top-K with no
sync; run() does the single stream.synchronize() then finish_topk to
assemble XFeatResult. Device buffers are capacity top_k; the valid count
is scores.len() — bound all device-buffer access by it.
Matching lives in a separate matching::Matcher (module crates/vrt-xfeat/src/matching.rs),
decoupled from postproc but sharing the stream. Cosine similarity (descriptors
are L2-normalized, so dot = cosine), mutual nearest-neighbor via two calls of one
tiled argmax kernel (xfeat_match_argmax, one thread per query, candidates tiled
through shared memory), min-similarity cutoff. VPI-style:
submit(Descriptors, Descriptors, cossim, &mut MatchResult) (async) → sync →
MatchResult::pairs().
The matcher is not XFeat-only. Matcher::new compiles for XFeat's 64-D descriptors;
Matcher::with_dim(stream, dim) compiles for 128-D as well, which is what matches
vrt-raco-aliked's ALIKED descriptors without LightGlue. The width is an NVRTC
compile-time constant, not a runtime argument, so a Matcher only ever handles the one
width it was built for.
That is why descriptors are passed as Descriptors::new(buf, count, dim) rather than a
bare slice and a count. The width cannot be recovered from the buffer — XFeatResult
allocates for top_k and RaCoAlikedResult for k, so both are longer than
count * dim in normal use, and no length arithmetic separates 3072x128 floats from
6144x64. Handing a 128-D set to a 64-D matcher does not crash; it strides the buffer and
returns plausible-looking matches. submit rejects a width that is not the compiled one,
along with a buffer too short for its count and a count over the output capacity — all
three as typed XFeatErrors, not debug assertions.
cargo test -p vrt-xfeat --release -- --ignored runs
gpu_match_agrees_with_cpu_reference (64-D), gpu_match_128d_agrees_with_cpu_reference
(128-D, plus the width/capacity/empty-side rejections) and
gpu_topk_selects_correct_keypoints. gpu_match_kernel_only_timing fails on JetPack 6
with Missing symbol cuEventElapsedTime_v2 — the driver exports only
cuEventElapsedTime, so cudarc's lookup cannot resolve; environmental, not a regression./255 happened exactly once (not zero, not twice).