Upload README.md with huggingface_hub
Browse files
README.md
ADDED
|
@@ -0,0 +1,17 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
tags: [coreml, face-recognition, arcface]
|
| 4 |
+
---
|
| 5 |
+
# ArcFace LResNet100E-IR — CoreML (fp16)
|
| 6 |
+
|
| 7 |
+
CoreML conversion of [arcfaceresnet100-8.onnx](https://github.com/onnx/models/tree/main/validated/vision/body_analysis/arcface)
|
| 8 |
+
from the ONNX Model Zoo (Apache-2.0), for on-device face embedding.
|
| 9 |
+
|
| 10 |
+
- Input: `faceImage` — 112x112 RGB image (raw 0-255 pixels; BN at graph input)
|
| 11 |
+
- Output: `embedding` — 512-d float vector (L2-normalize before use)
|
| 12 |
+
- mlprogram / fp16, iOS 17+
|
| 13 |
+
- Conversion: onnx (opset 8->13) -> onnx2torch -> coremltools.
|
| 14 |
+
Numerical parity vs onnxruntime verified (cosine 1.0 on real faces).
|
| 15 |
+
|
| 16 |
+
Packaged as `.mlpackage.tar.gz` for on-device compilation (mlprogram models
|
| 17 |
+
cannot ship as a single .mlmodel file).
|