FocsleFo’c’sle
IF

ArcFace-Edge

3%
by InsightFace

Reference face embedding for access control and retail re-ID.

Re-IDMITINT8FP16faceembeddingbiometric
218K downloads 16K deploymentsUpdated Jan 19, 2028
Headline:4.2ms · Hailo-8 · INT8
OverviewBenchmarks7Sim ResultsDeploy8FilesDiscussion23

About this model

Reference face embedding for access control and retail re-ID.

Authored by insightface. Curated into the Fo’c’sle reference set on 2028-01-19. All cross-chip benchmarks below were collected in matched-pair runs in the HIL lab using the same input pipeline, same upstream preprocessing, and the same downstream consumer. See the methodology page for the full protocol.

Task
Re-ID
Parameters
24.8 M
Benchmarked on
7 chips
Deployments
16K

Architecture

Embedding + similarity
Inferred from upstream weights · simplified
CropBackbonePoolingMargin headEmbedding

Headline benchmarks

QQualcomm QCS8550INT8
13.4ms p50
75 FPS98.5% acc11.8 W
AAmbarella CV72INT8
21.7ms p50
46 FPS97.2% acc5.0 W
πRaspberry Pi 5 + Hailo HATINT8
23.2ms p50
43 FPS98.7% acc7.6 W

Training data

Pretrained on the upstream maintainer’s released checkpoint. Edge-distillation pass uses 2.4M frames from the Fo’c’sle distillation corpus (consented public data + opt-in publisher contributions). Quantization-aware fine-tune uses 320K calibration samples drawn from the target task’s eval domain.

  • Pretraining corpus: upstream maintainer release
  • Distillation corpus: 2,400,000 frames
  • Calibration set: 320,000 samples (per task)
  • Eval set: standard benchmark + matched-pair HIL runs