Files
lightcraft/docs
Andrew FrahnandClaude Opus 5.5 f625c74c7a Faces: whole networks on the GPU, batched faces, CPU-only switch, paced GPU use
Activations stay on the device between layers (net/program.rs lowers the graph; every
operator is a WGSL kernel, one submission and read-back per run); faces of a photo go
in batches of up to 16; Settings > Faces shows each model's device and a CPU-only
switch (faces.gpu); the scan's pace limits GPU use; the GPU check is an activity row.
YuNet 58 -> 21 ms, SFace 40 -> 9.6 ms, AuraFace 295 -> 71 ms (heavily loaded machine).

Sources: own design.

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
2026-10-14 19:43:44 +11:00
..
2026-10-13 11:05:53 +07:00
2026-10-13 11:02:58 +07:00
2026-10-13 11:05:47 +07:00
2026-10-13 11:05:54 +07:00
2026-10-13 02:05:14 -04:00
2026-10-13 11:01:54 +07:00
2026-10-09 20:56:56 -04:00
2026-10-13 02:01:30 -04:00
2026-10-11 12:03:53 +00:00
2026-10-13 11:05:53 +07:00
2026-10-11 12:40:55 +00:00
2026-10-13 02:05:14 -04:00
2026-10-13 11:01:49 +07:00