mobilint/YOLO12m-face

YOLO12m_face

aries-rb/ holds the MXQ reproduced from the published arxiv folder (https://github.com/mobilint/mblt-model-arxiv/tree/main/models/face_detection/YOLO12m_face). It was compiled by qbcompiler 1.3.0 in the image mobilint/qbcompiler:1.3-cuda12.8.1-ubuntu22.04 (sha256:6ad69aaa9581487ac6a1864b88d4c2179c7a4a486a3ab1ac376dbdfe763415e7) with --inference-scheme single, so it supports the core mode(s) single only.

It was evaluated with qbruntime 1.5.0 in async inference mode on widerface (3226 samples): hard_ap 0.84861, easy_ap 0.95543, medium_ap 0.94516. Only the evaluated core mode was measured.

SHA-256: 6661c687cab80570e40526633a78c272fd55165af709d6790fe6f854f2e1f1a0 (28094750 bytes). Verify a download with sha256sum <file>.mxq and compare the digest with this value and with aries-rb/metadata.json.

aries-rb/best_result.json contains legacy build values and is retained for mblt-vision-python compile defaults. The legacy aries/ directory was removed; mblt-model-zoo <= 2.12 must pin an older revision.

{
  "schema_version": 2,
  "model": "YOLO12m_face",
  "variant": "bias-corr-taylor2-postact",
  "qbcompiler": "1.3.0",
  "qbruntime": "1.5.0",
  "compile": {
    "image_tag": "1.3-cuda12.8.1-ubuntu22.04",
    "image_digest": "sha256:6ad69aaa9581487ac6a1864b88d4c2179c7a4a486a3ab1ac376dbdfe763415e7",
    "inference_scheme": "single",
    "device": "gpu"
  },
  "evaluation": {
    "dataset": "widerface",
    "sample_count": 3226,
    "primary": {
      "hard_ap": 0.84861
    },
    "secondary": {
      "easy_ap": 0.95543,
      "medium_ap": 0.94516
    },
    "protocol": null,
    "eval_data_identity": null,
    "identity_audit": "absent",
    "inference_mode": "async",
    "image_tag": "1.3-cpu-ubuntu22.04",
    "image_digest": "sha256:7c25b894fc0807acdf27f15c61b910b96dc66d3f5d97e04b595b2d7a8008e77c"
  },
  "retention_percent": {
    "hard_ap": 99.12394435294529
  },
  "content_sha256": "3179a4244cd56410eec290c02746163e2a4753653dd66f45285086d61ba33cb4",
  "attempt_id": "srcmodes10-yolo12m-face-20261010t175053",
  "superseded_attempt_ids": [],
  "sha256": "6661c687cab80570e40526633a78c272fd55165af709d6790fe6f854f2e1f1a0",
  "format": "MXQv7",
  "supported_modes": [
    "single"
  ],
  "size": 28094750,
  "ops_commits": {
    "selection": "66d2f03be3d5af292a3bd707a630737ad5631488"
  },
  "arxiv_folder": "https://github.com/mobilint/mblt-model-arxiv/tree/main/models/face_detection/YOLO12m_face",
  "publication": {
    "status": "public",
    "reason": "approved"
  }
}