compiled artefacts from mac_arm64 (cpu)

.gitattributes

@@ -27,3 +27,5 @@ code-daemon-denoise-v1-s_win_x64_trt11.3_sm_75.engine filter=lfs diff=lfs merge=
 code-daemon-denoise-v1-s_win_x64_trt11.3_sm_80.engine filter=lfs diff=lfs merge=lfs -text
 code-daemon-denoise-v1-s_win_x64_trt11.3_sm_89.engine filter=lfs diff=lfs merge=lfs -text
 code-daemon-denoise-v1-s_win_x64_trt11.3_sm_86.engine filter=lfs diff=lfs merge=lfs -text
+coreml_ane/embed.mlpackage/Data/com.apple.CoreML/model.mlmodel filter=lfs diff=lfs merge=lfs -text
+coreml_ane/embed.mlpackage/Data/com.apple.CoreML/weights/weight.bin filter=lfs diff=lfs merge=lfs -text

README.md

@@ -253,6 +253,28 @@ the vocabulary prune and is no longer published; they are corrected rather than
 
 ---
 
+## Apple Neural Engine (Core ML)
+
+`coreml_ane/embed.mlpackage/` is a Core ML **multifunction** package that runs this
+denoising encoder on the Apple Neural Engine. It is a bundle, not a file — four
+entries (`Manifest.json`, `shapes.json`, `Data/com.apple.CoreML/model.mlmodel`,
+`Data/com.apple.CoreML/weights/weight.bin`) that must keep their relative paths.
+
+**One compiled function per shape**, named `b<batch>_s<seq>`: `b64 s40`.
+Fixed shapes are not a simplification. `ct.EnumeratedShapes` converts and runs, and
+measures 233 emb/s against 2 873 on the same encoder, because the dynamic ops it
+injects push the graph off the Neural Engine. `shapes.json` lists what was compiled,
+so a caller can ask instead of assuming.
+
+Weights are fp16. The package is built from the MLX safetensors beside it, so a model
+is ANE-ready exactly when it is MLX-ready — there is no second set of source weights.
+Load it with `MLComputeUnits.cpuAndNeuralEngine`: plain `.all` lets Core ML place the
+graph on the GPU instead, which measured 809 emb/s against the ANE's 2 873.
+
+A shape the package does not carry is not an error — the caller is expected to fall
+back to the MLX graph, which takes any shape. That is what makes the fixed-shape
+package safe to ship alongside `model_gpu_mlx*/` rather than instead of it.
+
 ## 8. License & attribution
 
 The encoder weights are **[`intfloat/multilingual-e5-small`](https://huggingface.co/intfloat/multilingual-e5-small)**

coreml_ane/embed.mlpackage/Data/com.apple.CoreML/model.mlmodel

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:507a8e6cb868fe39d33f140d6f45c88968dfc29477cba3a7e74faf8250bdaa9e
+size 119033

coreml_ane/embed.mlpackage/Data/com.apple.CoreML/weights/weight.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f9058955c30a87e5a629277ead3b8a3158f1f51a3e6d48c7ed2ee193f56c881c
+size 236568192

coreml_ane/embed.mlpackage/Manifest.json

@@ -0,0 +1,18 @@
+{
+    "fileFormatVersion": "1.0.0",
+    "itemInfoEntries": {
+        "7B3E00EF-CFC5-4864-A5AA-21036D4E8AA5": {
+            "author": "com.apple.CoreML",
+            "description": "CoreML Model Specification",
+            "name": "model.mlmodel",
+            "path": "com.apple.CoreML/model.mlmodel"
+        },
+        "C5D2218E-6943-4201-80DE-35422B300978": {
+            "author": "com.apple.CoreML",
+            "description": "CoreML Model Weights",
+            "name": "weights",
+            "path": "com.apple.CoreML/weights"
+        }
+    },
+    "rootModelIdentifier": "7B3E00EF-CFC5-4864-A5AA-21036D4E8AA5"
+}

coreml_ane/embed.mlpackage/shapes.json

@@ -0,0 +1,12 @@
+{
+  "head": "mean",
+  "output_dim": 384,
+  "per_token": false,
+  "shapes": [
+    {
+      "batch": 64,
+      "seq": 40,
+      "function": "b64_s40"
+    }
+  ]
+}

manifest.json

@@ -5,7 +5,6 @@
   "quantization": "int8",
   "targets": [
     "openvino",
-    "tensorrt",
     "tvm"
   ],
   "runtime_tags": {
@@ -14,6 +13,6 @@
     "tvm": "tvm0.25",
     "mlx_dir": "model_gpu_mlx0.32"
   },
-  "compiled_at": "2026-09-17T21:53:32Z",
+  "compiled_at": "2026-09-24T02:26:30Z",
   "compiled_by": "models/_compile"
 }