face analysis benchmarks on edge AI hardware
94 published rows. Each table compares only results with the same model-size class and batch size. Search this workload · JSON.
Fastest in the best-covered like-for-like group (mid-size CNN, batch 1): MemryX MX3 M.2 module at 7221 FPS C 2026-07-15 (published). 3 groups, 11 platforms.
Show data — 3 like-for-like tables
mid-size CNN · batch 1 (11 platforms)
✓ Like-for-like: model-size class mid-size CNN, batch 1. Model, precision and stack may still differ.
| Platform | Result | Model · precision · stack | Grade | Power | Source |
|---|---|---|---|---|---|
| MemryX MX3 M.2 module | 7221 FPS throughput · b=1 | YuNet not stated (MX3 activations BF16 / weights 4-8-16 bit per vendor) · MemryX SDK 2.2 (closed-loop compiled DFP) · mid-size CNN | C | unknown power n/a no power published | MemryX Model Explorer (SDK 2.2) 2026-07-15 (published) |
| Axelera Metis Embedded 113m M.2 card | 1285 FPS throughput · b=1 | FaceNet - InceptionResnetV1 INT8 · Voyager SDK (GStreamer pipeline) · mid-size CNN | C | unknown power n/a no power published | Axelera Voyager SDK model zoo (docs/reference/models/model-zoo.md) 2026-10-15 (page last edited) |
| Axelera Metis Edge 130p PCIe card | 1226 FPS throughput · b=1 | FaceNet - InceptionResnetV1 INT8 · Voyager SDK (GStreamer pipeline) · mid-size CNN | C | unknown power n/a no power published | Axelera Voyager SDK model zoo (docs/reference/models/model-zoo.md) 2026-10-15 (page last edited) |
| Rockchip RK3576 (NPU, single core of 2) | 54.60 FPS throughput · b=1 | RetinaFace_resnet50_320 INT8 · RKNN-Toolkit2 / RKNPU (rknn_model_zoo) · mid-size CNN | C | unknown power n/a no power published | Rockchip rknn_model_zoo README benchmark table 2026-10-13 (page last edited) |
| Rockchip RK3588 (NPU, single core of 3) | 50.00 FPS throughput · b=1 | RetinaFace_resnet50_320 INT8 · RKNN-Toolkit2 / RKNPU (rknn_model_zoo) · mid-size CNN | C | unknown power n/a no power published | Rockchip rknn_model_zoo README benchmark table 2026-10-13 (page last edited) |
| Rockchip RV1126B (NPU) | 46.50 FPS throughput · b=1 | RetinaFace_resnet50_320 INT8 · RKNN-Toolkit2 / RKNPU (rknn_model_zoo) · mid-size CNN | C | unknown power n/a no power published | Rockchip rknn_model_zoo README benchmark table 2026-10-13 (page last edited) |
| Rockchip RK3562 (NPU) | 27.10 FPS throughput · b=1 | RetinaFace_resnet50_320 INT8 · RKNN-Toolkit2 / RKNPU (rknn_model_zoo) · mid-size CNN | C | unknown power n/a no power published | Rockchip rknn_model_zoo README benchmark table 2026-10-13 (page last edited) |
| Rockchip RK1808 (NPU) | 24.60 FPS throughput · b=1 | RetinaFace_resnet50_320 INT8 · RKNN-Toolkit2 / RKNPU (rknn_model_zoo) · mid-size CNN | C | unknown power n/a no power published | Rockchip rknn_model_zoo README benchmark table 2026-10-13 (page last edited) |
| Rockchip RV1126 (NPU) | 20.80 FPS throughput · b=1 | RetinaFace_resnet50_320 INT8 · RKNN-Toolkit2 / RKNPU (rknn_model_zoo) · mid-size CNN | C | unknown power n/a no power published | Rockchip rknn_model_zoo README benchmark table 2026-10-13 (page last edited) |
| Rockchip RK3566/RK3568 (NPU) | 18.40 FPS throughput · b=1 | RetinaFace_resnet50_320 INT8 · RKNN-Toolkit2 / RKNPU (rknn_model_zoo) · mid-size CNN | C | unknown power n/a no power published | Rockchip rknn_model_zoo README benchmark table 2026-10-13 (page last edited) |
| Rockchip RV1109 (NPU) | 14.60 FPS throughput · b=1 | RetinaFace_resnet50_320 INT8 · RKNN-Toolkit2 / RKNPU (rknn_model_zoo) · mid-size CNN | C | unknown power n/a no power published | Rockchip rknn_model_zoo README benchmark table 2026-10-13 (page last edited) |
mobile-class CNN · batch 1 (1 platforms)
✓ Like-for-like: model-size class mobile-class CNN, batch 1. Model, precision and stack may still differ.
| Platform | Result | Model · precision · stack | Grade | Power | Source |
|---|---|---|---|---|---|
| MemryX MX3 M.2 module | 19560 FPS throughput · b=1 | MediaPipe-iris-landk not stated (MX3 activations BF16 / weights 4-8-16 bit per vendor) · MemryX SDK 2.2 (closed-loop compiled DFP) · mobile-class CNN | C | unknown power n/a no power published | MemryX Model Explorer (SDK 2.2) 2026-07-15 (published) |
unknown · batch 1 (9 platforms)
✓ Like-for-like: model-size class unknown, batch 1 (size or batch not stated by source — unverified). Model, precision and stack may still differ.
| Platform | Result | Model · precision · stack | Grade | Power | Source |
|---|---|---|---|---|---|
| MemryX MX3 M.2 module | 37925 FPS throughput · b=1 | Open-closed-eye-0001 not stated (MX3 activations BF16 / weights 4-8-16 bit per vendor) · MemryX SDK 2.2 (closed-loop compiled DFP) · unknown | C | unknown power n/a no power published | MemryX Model Explorer (SDK 2.2) 2026-07-15 (published) |
| Rockchip RK3576 (NPU, single core of 2) | 502 FPS throughput · b=1 | RetinaFace_mobile320 INT8 · RKNN-Toolkit2 / RKNPU (rknn_model_zoo) · unknown | C | unknown power n/a no power published | Rockchip rknn_model_zoo README benchmark table 2026-10-13 (page last edited) |
| Rockchip RV1126B (NPU) | 452 FPS throughput · b=1 | RetinaFace_mobile320 INT8 · RKNN-Toolkit2 / RKNPU (rknn_model_zoo) · unknown | C | unknown power n/a no power published | Rockchip rknn_model_zoo README benchmark table 2026-10-13 (page last edited) |
| Rockchip RK3562 (NPU) | 353 FPS throughput · b=1 | RetinaFace_mobile320 INT8 · RKNN-Toolkit2 / RKNPU (rknn_model_zoo) · unknown | C | unknown power n/a no power published | Rockchip rknn_model_zoo README benchmark table 2026-10-13 (page last edited) |
| Rockchip RK3588 (NPU, single core of 3) | 229 FPS throughput · b=1 | RetinaFace_mobile320 INT8 · RKNN-Toolkit2 / RKNPU (rknn_model_zoo) · unknown | C | unknown power n/a no power published | Rockchip rknn_model_zoo README benchmark table 2026-10-13 (page last edited) |
| Rockchip RV1126 (NPU) | 213 FPS throughput · b=1 | RetinaFace_mobile320 INT8 · RKNN-Toolkit2 / RKNPU (rknn_model_zoo) · unknown | C | unknown power n/a no power published | Rockchip rknn_model_zoo README benchmark table 2026-10-13 (page last edited) |
| Rockchip RK1808 (NPU) | 199 FPS throughput · b=1 | RetinaFace_mobile320 INT8 · RKNN-Toolkit2 / RKNPU (rknn_model_zoo) · unknown | C | unknown power n/a no power published | Rockchip rknn_model_zoo README benchmark table 2026-10-13 (page last edited) |
| Rockchip RK3566/RK3568 (NPU) | 149 FPS throughput · b=1 | RetinaFace_mobile320 INT8 · RKNN-Toolkit2 / RKNPU (rknn_model_zoo) · unknown | C | unknown power n/a no power published | Rockchip rknn_model_zoo README benchmark table 2026-10-13 (page last edited) |
| Rockchip RV1109 (NPU) | 145 FPS throughput · b=1 | RetinaFace_mobile320 INT8 · RKNN-Toolkit2 / RKNPU (rknn_model_zoo) · unknown | C | unknown power n/a no power published | Rockchip rknn_model_zoo README benchmark table 2026-10-13 (page last edited) |
How to read these numbers
Every number is labelled. Confidence: A MLPerf closed, audited · B documented third-party method · C community / review · D vendor-published, not independently verified · E estimate. Power basis: measured with the result · spec vendor spec/envelope · power n/a not published.
Machine-readable: /find.json (see llms.txt); full tables are in “Show data”. Confidence grades: A audited MLPerf closed · B documented third-party · C community · D vendor-published · E estimate. Unknown values are shown as unknown, never estimated. Generated 2026-10-16 from public sources.