mirror of
https://github.com/firestar5683/StarPilot.git
synced 2026-11-12 23:43:44 +08:00
firestar5683
58427d61d9
Select qualified speed-limit sources, preserve acceptance and override decisions, and publish distinct control and display coordinates. Carry accepted and pending source identity with each displayed decision. Co-authored-by: firestarsdog <229254897+firestarsdog@users.noreply.github.com>
865 B
865 B
These two ONNX files are copied byte-for-byte from the frozen StarPilot
678af783 baseline under starpilot/assets/vision_models/:
| File | Bytes | SHA-256 |
|---|---|---|
speed_limit_us_detector.onnx |
9,605,948 | 82408b68c79c269296f0af942130c5383cace4ee06c78e2a4690e8488720116a |
speed_limit_us_value_classifier.onnx |
6,229,930 | 07c6696e530eb940d2757d5849b4bc0f1d785cda704e5296e18c0a94959f30a5 |
The frozen tree does not identify the training set, original model publisher, or an asset-specific redistribution license. These facts remain unresolved before production/device distribution. The current service is development opt-in.
On the local OpenCV 4.13 host, the detector accepts a 256×256 tensor and emits
(1, 5, 1344); the classifier accepts 128×128 and emits (1, 19). Device
inference compatibility has not been established.