Move the panorama keypoint detector onto the engine, and probe with a detector

XFeat's two exports are a Keypoints role now; the crate no longer names
tract, and the app compiles TensorRT engines for both ahead of the
first merge. The probe picks the smallest *detector* rather than the
smallest file: the tablet's first run chose the 112 KB eye classifier,
which has no int8 form, and reported the Hexagon as failed for want of
one.

Cargo.lock

@@ -1550,12 +1550,12 @@ name = "dr-pano"
 version = "0.12.2"
 dependencies = [
  "dr-decode",
+ "dr-inference-engine",
  "dr-types",
  "env_logger",
  "log",
  "ndarray",
  "ort",
- "ort-tract",
  "thiserror 2.0.20",
 ]