Search results

Assess location precision for the unassessed candidate cities — the gate every route to 500k depends on

documentationImprovements or additions to documentationImprovements or additions to documentation

Status: Open.

#96 In ProjectSidewalk/RampNet;

· jonfroehlich opened on Jul 31, 2026

Attribute any RampNet-vs-YOLO gap: backbone, output representation, and training budget are confounded in #51

Status: Open.

#95 In ProjectSidewalk/RampNet;

· jonfroehlich opened on Jul 31, 2026

Tune the YOLO baseline's peak LR / warmup before reporting it — the collapse is an untuned default (#51)

Status: Open.

#90 In ProjectSidewalk/RampNet;

· jonfroehlich opened on Jul 30, 2026

RampNet 2.0: measurement, condition inference, and tagging — we already collect the supervision and discard it at ingest

Status: Open.

#86 In ProjectSidewalk/RampNet;

· jonfroehlich opened on Jul 29, 2026

Aerial orthoimagery curb-ramp detection already exists in production (34k ramps at 1-inch GSD): treat it as an uncorrelated arbiter, not a competitor

Status: Open.

#85 In ProjectSidewalk/RampNet;

· jonfroehlich opened on Jul 29, 2026

Stage 2 trains for 1 epoch at constant LR (~9.4k steps): measure the epoch curve, and with it the auto-label ceiling

Status: Open.

#84 In ProjectSidewalk/RampNet;

· jonfroehlich opened on Jul 29, 2026

Extend RampNet output from points to boxes/polygons (point-prompted SAM2; CenterNet-style size head; segmentation)

enhancementNew feature or requestNew feature or request

Status: Open.

#83 In ProjectSidewalk/RampNet;

· jonfroehlich opened on Jul 29, 2026

Training augmentation is a single horizontal flip: measure downscale / photometric jitter as a transfer lever (downstream of #84)

Status: Open.

#82 In ProjectSidewalk/RampNet;

· jonfroehlich opened on Jul 29, 2026

Second rater for Budapest District V: the rubric is now confounded with the benchmark's only ranking inversion

Status: Open.

#74 In ProjectSidewalk/RampNet;

· jonfroehlich opened on Jul 28, 2026

Document the YOLOv11-pano training instability as a reported caveat

documentationImprovements or additions to documentationImprovements or additions to documentation

enhancementNew feature or requestNew feature or request

Status: Open.

#72 In ProjectSidewalk/RampNet;

· jonfroehlich opened on Jul 28, 2026

Rerun the collapsed YOLOv11 pano baselines with a stabilized (larger physical batch) config

enhancementNew feature or requestNew feature or request

Status: Open.

#70 In ProjectSidewalk/RampNet;

· jonfroehlich opened on Jul 28, 2026

Would 2-3M training labels buy recall? Measure the scaling curve and the label ceiling before a sourcing campaign

enhancementNew feature or requestNew feature or request

Status: Open.

#59 In ProjectSidewalk/RampNet;

· jonfroehlich opened on Jul 25, 2026