Upload README.md with huggingface_hub

README.md

@@ -8,7 +8,7 @@ tags:
 - image-segmentation
 - portrait-matting
 ---
-# modnet_portrait_matting — ExecuTorch XNNPACK
+# modnet_portrait_matting — ExecuTorch
 
 - **Source**: ZHKKKe/MODNet + DavG25/modnet-pretrained-models ckpt
 - **License**: Apache-2.0
@@ -19,15 +19,23 @@ tags:
 
 All variants take and return fp32 tensors — swap the `.pte` file, keep your app code.
 
-| precision | file | size (MB) | parity vs fp32 eager (worst corr) | Mac median (ms)* |
+| build | file | size (MB) | parity vs fp32 eager (worst corr) | Mac median (ms)* |
 |-----------|------|-----------|------------------------------------|------------------|
 | fp32 | `modnet_portrait_matting_xnnpack_fp32.pte` | 26.1 | 1.000000 | 64.9 |
 | fp16 | `modnet_portrait_matting_xnnpack_fp16.pte` | 24.4 | 1.000000 | 107.8 |
+| Core ML (fp16, iOS) | `modnet_portrait_matting_coreml_all.pte` | 13.8 | 0.999997 | 6.8 |
+
+
+The Core ML build is the same graph lowered to Apple's Neural Engine instead of
+XNNPACK, which is CPU-only. On an iPhone 17 Pro, Depth-Anything-V2-Small runs
+500.8 ms through XNNPACK and 42.7 ms through Core ML, at half the file size. It
+computes in fp16 and is iOS-only; the XNNPACK files stay the portable option and
+are what runs on Android.
 
 \*Mac arm64, single process, median of 10 — a reference point for relative cost
 only, not a device number (torch eager fp32 on the same machine: 120.1 ms).
 
-### Precisions that did not earn a slot
+### Builds that did not earn a slot
 
 - **int8 is not shipped**: measured in the units that matter for this model — mask IoU at 0.5: median 0.9864 over 10 real images, worst 0.5970.
 
@@ -44,5 +52,5 @@ XNNPACK delegate coverage (fp32): 94.7% (302/319 ops); ops left on the portable
 
 ## Conversion
 
-torch.export -> to_edge_transform_and_lower(XnnpackPartitioner) -> .pte
+torch.export -> to_edge_transform_and_lower(partitioner) -> .pte
 (conversion scripts: [executorch-models](https://github.com/john-rocky/executorch-models))