Re-export under the get_model_schema contract; restructure to the MODEL_SPEC layout

.gitattributes

@@ -36,3 +36,6 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 PP-OCRv6_coreml.pte filter=lfs diff=lfs merge=lfs -text
 PP-OCRv6_vulkan.pte filter=lfs diff=lfs merge=lfs -text
 PP-OCRv6_xnnpack.pte filter=lfs diff=lfs merge=lfs -text
+coreml/pp_ocrv6_coreml_int8.pte filter=lfs diff=lfs merge=lfs -text
+vulkan/pp_ocrv6_vulkan_fp16.pte filter=lfs diff=lfs merge=lfs -text
+xnnpack/pp_ocrv6_xnnpack_fp32.pte filter=lfs diff=lfs merge=lfs -text

README.md

@@ -19,31 +19,44 @@ per-size method buckets). The `.pte` is a pure tensor→tensor function; all pre
 (resize, normalize, DBNet box decode, perspective crop, CTC decode) is the client's job and is
 driven by `config.json`. One model covers **all languages** (18 709-entry multilingual charset).
 
+## Repository layout
+
+```
+<backend>/config.json                    # per-backend spec (see the schema link inside)
+<backend>/pp_ocrv6_<backend>_<precision>.pte
+charset.txt                              # 18 709 entries; charset[i] -> logit i+1, blank = 0
+```
+
 ## Methods & I/O contract
 
 | method | input | output |
 |---|---|---|
 | `detect` (DBNet) | `[1,3,H,W]` f32 RGB, **ImageNet-normalized by the client**: `(x/255 − mean)/std`, `mean=[0.485,0.456,0.406]`, `std=[0.229,0.224,0.225]` | `[1,1,H,W]` probability map (sigmoid baked) |
-| `recognize` (SVTR) | `[1,3,48,W]` f32 RGB, client-normalized `(x/255 − 0.5)/0.5` | `[1,W/8,18709+1]` probs (softmax baked); `charset[i]` → logit `i+1`, blank = 0 |
+| `recognize` (SVTR) | `[1,3,48,W]` f32 RGB, client-normalized `(x/255 − 0.5)/0.5` | `[1,W/8,18710]` probs (softmax baked); `charset[i]` → logit `i+1`, blank = 0 |
 
 **Nothing is baked for input normalization** — the client normalizes before calling.
 Note the two methods use *different* norms (ImageNet for detect, `0.5/0.5` for recognize).
 
-## Shape discovery (companion methods)
-
-Every method carries exactly **one** no-arg discovery companion:
+## Shape discovery (`get_model_schema`)
 
-- `get_dynamic_dims_<method>` — dynamic method. Returns one `int32 [rank, 3]` tensor per
-  tensor input; each row is `[min, max, step]` (static dims are `[n, n, 1]`). Any conforming
-  shape is valid.
-- `get_enum_shapes_<method>` — enumerated method. Returns one `int32 [N, rank]` tensor per
-  tensor input; each row is a complete legal shape (cross-dimension coupling is exact —
-  a listed `1280×640` does **not** imply `640×1280`). Snap inputs to the nearest row.
+Every `.pte` exports one no-arg constant method, **`get_model_schema`**, returning a JSON
+`ModelSpec` string: per method, the input and output parameter specs (dtype plus a domain per
+dimension — `constant`, `range` with `{min, max, step}`, or `enum` with explicit `choices`) and
+the runtime constraints the method declares over its dimensions. Methods absent from the JSON
+are fully static and described by ExecuTorch's own `MethodMeta`. The schema is validated
+against
+[`config.schema.json`'s `ModelSpec`](https://huggingface.co/software-mansion/react-native-executorch-spec)
+counterpart in the library at load time. The older `get_dynamic_dims_<m>` / `get_enum_shapes_<m>`
+companion methods are **gone** — everything they carried now lives in this one document.
 
-| backend | `detect` | `recognize` |
+| backend | `detect` H, W | `recognize` W (H fixed 48) |
 |---|---|---|
-| `xnnpack`, `vulkan` | `get_dynamic_dims_detect` → H, W ∈ `[640, 1280]` step 32 | `get_dynamic_dims_recognize` → W ∈ `[160, 1280]` step 8 (H fixed 48) |
-| `coreml` | `get_enum_shapes_detect` → `640²`, `960²`, `1280²`, `1280×640` | `get_enum_shapes_recognize` → widths `160, 320, 480, 640, 1280` |
+| `xnnpack`, `vulkan` | `range` `[640, 1280]` step 32 | `range` `[160, 1280]` step 8 |
+| `coreml` | `enum` `640, 960, 1280` per dimension | `enum` `160, 320, 480, 640, 1280` |
+
+`recognize` additionally declares a **linear runtime constraint** tying its input width to its
+CTC timestep count — `width = 8 × timesteps + 0` — so a client can size the probs output for
+whatever width it picks instead of inferring a ratio.
 
 ## Backends
 
@@ -77,3 +90,6 @@ the compatibility note in the
 [ExecuTorch GitHub repository](https://github.com/pytorch/executorch/blob/main/runtime/COMPATIBILITY.md).
 If you work with React Native ExecuTorch, the library constants guarantee compatibility with the
 runtime used behind the scenes.
+
+These models were exported with ExecuTorch 1.3.1 and **no forward compatibility** is
+guaranteed; older runtimes may not load them.

config.json

@@ -1,20 +1,3 @@
 {
-  "schemaVersion": 1,
-  "detectorKind": "dbnet",
-  "buckets": {
-    "detect": [
-      640,
-      960,
-      1280
-    ],
-    "recognize": [
-      160,
-      320,
-      480,
-      640,
-      1280
-    ]
-  },
-  "charsetUrl": "charset.txt",
-  "dropScore": 0.5
-}
+  "modelName": "pp-ocrv6"
+}

coreml/config.json

@@ -0,0 +1,69 @@
+{
+  "$schema": "https://huggingface.co/software-mansion/react-native-executorch-spec/resolve/main/config.schema.json",
+  "model": "pp_ocrv6",
+  "family": "paddleocr",
+  "capabilities": [
+    "text-detection",
+    "text-recognition"
+  ],
+  "backend": "coreml",
+  "license": "apache-2.0",
+  "variants": [
+    {
+      "file": "pp_ocrv6_coreml_int8.pte",
+      "precision": "int8",
+      "quantized": true,
+      "default": true,
+      "size_bytes": 8269910,
+      "methods": {
+        "detect": {
+          "inputs": [
+            {
+              "shape": [
+                1,
+                3,
+                -1,
+                -1
+              ],
+              "dtype": "float32"
+            }
+          ],
+          "outputs": [
+            {
+              "shape": [
+                1,
+                1,
+                -1,
+                -1
+              ],
+              "dtype": "float32"
+            }
+          ]
+        },
+        "recognize": {
+          "inputs": [
+            {
+              "shape": [
+                1,
+                3,
+                48,
+                -1
+              ],
+              "dtype": "float32"
+            }
+          ],
+          "outputs": [
+            {
+              "shape": [
+                1,
+                -1,
+                18710
+              ],
+              "dtype": "float32"
+            }
+          ]
+        }
+      }
+    }
+  ]
+}

coreml/pp_ocrv6_coreml_int8.pte

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:170309543ff4f519b304aa40be96be36224ed92b24760fade21af9909df771d3
+size 8269910

vulkan/config.json

@@ -0,0 +1,69 @@
+{
+  "$schema": "https://huggingface.co/software-mansion/react-native-executorch-spec/resolve/main/config.schema.json",
+  "model": "pp_ocrv6",
+  "family": "paddleocr",
+  "capabilities": [
+    "text-detection",
+    "text-recognition"
+  ],
+  "backend": "vulkan",
+  "license": "apache-2.0",
+  "variants": [
+    {
+      "file": "pp_ocrv6_vulkan_fp16.pte",
+      "precision": "fp16",
+      "quantized": false,
+      "default": true,
+      "size_bytes": 26176856,
+      "methods": {
+        "detect": {
+          "inputs": [
+            {
+              "shape": [
+                1,
+                3,
+                -1,
+                -1
+              ],
+              "dtype": "float32"
+            }
+          ],
+          "outputs": [
+            {
+              "shape": [
+                1,
+                1,
+                -1,
+                -1
+              ],
+              "dtype": "float32"
+            }
+          ]
+        },
+        "recognize": {
+          "inputs": [
+            {
+              "shape": [
+                1,
+                3,
+                48,
+                -1
+              ],
+              "dtype": "float32"
+            }
+          ],
+          "outputs": [
+            {
+              "shape": [
+                1,
+                -1,
+                18710
+              ],
+              "dtype": "float32"
+            }
+          ]
+        }
+      }
+    }
+  ]
+}

vulkan/pp_ocrv6_vulkan_fp16.pte

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:71a26547bb71bbc66d829648183ea112d4d5c1ec71295cea847e6c4d6ffad818
+size 26176856

xnnpack/config.json

@@ -0,0 +1,69 @@
+{
+  "$schema": "https://huggingface.co/software-mansion/react-native-executorch-spec/resolve/main/config.schema.json",
+  "model": "pp_ocrv6",
+  "family": "paddleocr",
+  "capabilities": [
+    "text-detection",
+    "text-recognition"
+  ],
+  "backend": "xnnpack",
+  "license": "apache-2.0",
+  "variants": [
+    {
+      "file": "pp_ocrv6_xnnpack_fp32.pte",
+      "precision": "fp32",
+      "quantized": false,
+      "default": true,
+      "size_bytes": 31066200,
+      "methods": {
+        "detect": {
+          "inputs": [
+            {
+              "shape": [
+                1,
+                3,
+                -1,
+                -1
+              ],
+              "dtype": "float32"
+            }
+          ],
+          "outputs": [
+            {
+              "shape": [
+                1,
+                1,
+                -1,
+                -1
+              ],
+              "dtype": "float32"
+            }
+          ]
+        },
+        "recognize": {
+          "inputs": [
+            {
+              "shape": [
+                1,
+                3,
+                48,
+                -1
+              ],
+              "dtype": "float32"
+            }
+          ],
+          "outputs": [
+            {
+              "shape": [
+                1,
+                -1,
+                18710
+              ],
+              "dtype": "float32"
+            }
+          ]
+        }
+      }
+    }
+  ]
+}

xnnpack/pp_ocrv6_xnnpack_fp32.pte

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fad0b9ed668021ac5c4afc4501ec59c0eb9cc1dccf8b5ed627ab0a51a34d30df
+size 31066200