| --- |
| license: bsd-3-clause |
| pipeline_tag: image-segmentation |
| library_name: executorch |
| --- |
| |
| # lraspp |
|
|
| This repository hosts the **lraspp** models exported for the |
| [React Native ExecuTorch](https://www.npmjs.com/package/react-native-executorch) |
| library as ExecuTorch `.pte` programs, ready to run on device. |
|
|
| Upstream model: [LRASPP](https://docs.pytorch.org/vision/main/models/lraspp.html) |
|
|
| ## Variants |
|
|
| | Path | Backend | Precision | |
| | --- | --- | --- | |
| | `coreml/lraspp_mobilenet_v3_large_coreml_fp16.pte` | coreml | fp16 | |
| | `xnnpack/lraspp_mobilenet_v3_large_xnnpack_fp32.pte` | xnnpack | fp32 | |
| | `xnnpack/lraspp_mobilenet_v3_large_xnnpack_int8.pte` | xnnpack | int8 | |
| | `qnn/lraspp_mobilenet_v3_large_qnn_a8w8_v69.pte` | qnn | a8w8 | |
| | `qnn/lraspp_mobilenet_v3_large_qnn_a8w8_v73.pte` | qnn | a8w8 | |
| | `qnn/lraspp_mobilenet_v3_large_qnn_a8w8_v75.pte` | qnn | a8w8 | |
| | `qnn/lraspp_mobilenet_v3_large_qnn_a8w8_v79.pte` | qnn | a8w8 | |
| | `qnn/lraspp_mobilenet_v3_large_qnn_a8w8_v81.pte` | qnn | a8w8 | |
|
|
| ## QNN variants |
|
|
| A QNN `.pte` embeds an HTP context binary compiled for one Hexagon version, so |
| there is one file per version and the runtime picks the one matching the device. |
| These return a class index per pixel, `(1, 520, 520)` int32, rather than |
| `(1, 21, 520, 520)` logits. |
|
|
| | Hexagon | Example SoC | |
| | --- | --- | |
| | v69 | Snapdragon 8 Gen 1 | |
| | v73 | Snapdragon 8 Gen 2 | |
| | v75 | Snapdragon 8 Gen 3 | |
| | v79 | Snapdragon 8 Elite | |
| | v81 | Snapdragon 8 Elite Gen 5 | |
|
|
| ## Unshipped backends |
|
|
| A backend directory carrying a `NOTES.md` and no `.pte` was exported and |
| not published. Its note records what went wrong. |
|
|
| - `mlx` ([why](https://huggingface.co/software-mansion/react-native-executorch-lraspp/blob/main/mlx/NOTES.md)) |
|
|
| ## Repository structure |
|
|
| ``` |
| config.json 28 B |
| coreml/config.json 899 B |
| coreml/lraspp_mobilenet_v3_large_coreml_fp16.pte 6.6 MB |
| mlx/NOTES.md 2.0 kB |
| xnnpack/config.json 1.5 kB |
| xnnpack/lraspp_mobilenet_v3_large_xnnpack_fp32.pte 12.3 MB |
| xnnpack/lraspp_mobilenet_v3_large_xnnpack_int8.pte 3.4 MB |
| ``` |
|
|
| ## Compatibility |
|
|
| These files are published for the **ExecuTorch v1.4.1** runtime. ExecuTorch |
| gives no forward compatibility guarantee, so an older runtime may fail to load |
| them. |
|
|
| To use them in React Native ExecuTorch, pass the model constant shipped in the |
| library's model registry to the corresponding task pipeline. See the |
| [documentation](https://docs.swmansion.com/react-native-executorch/docs/fundamentals/downloading-models). |
|
|
| To load these files in your own ExecuTorch runtime, read the |
| [compatibility note](https://github.com/pytorch/executorch/blob/main/runtime/COMPATIBILITY.md) |
| first. |
|
|
