software-mansion/react-native-executorch-efficientnet-v2-s

1---2license: other3license_name: apache-license-2.04license_link: https://github.com/google/automl/blob/master/LICENSE5pipeline_tag: image-classification6library_name: executorch7---8 9# efficientnet-v2-s10 11This repository hosts the **efficientnet-v2-s** models exported for the12[React Native ExecuTorch](https://www.npmjs.com/package/react-native-executorch)13library as ExecuTorch `.pte` programs, ready to run on device.14 15Upstream model: [efficientnet_v2_s](https://pytorch.org/vision/0.20/models/generated/torchvision.models.efficientnet_v2_s.html#torchvision.models.efficientnet_v2_s)16 17## Variants18 19| Path | Backend | Precision |20| --- | --- | --- |21| `coreml/efficientnet_v2_s_coreml_fp16.pte` | coreml | fp16 |22| `xnnpack/efficientnet_v2_s_xnnpack_fp32.pte` | xnnpack | fp32 |23| `xnnpack/efficientnet_v2_s_xnnpack_int8.pte` | xnnpack | int8 |24 25## Unshipped backends26 27A backend directory carrying a `NOTES.md` and no `.pte` was exported and28not published. Its note records what went wrong.29 30- `mlx` ([why](https://huggingface.co/software-mansion/react-native-executorch-efficientnet-v2-s/blob/main/mlx/NOTES.md))31 32## Repository structure33 34```35config.json                                 39 B36coreml/config.json                          840 B37coreml/efficientnet_v2_s_coreml_fp16.pte    41.8 MB38mlx/NOTES.md                                2.5 kB39xnnpack/config.json                         1.4 kB40xnnpack/efficientnet_v2_s_xnnpack_fp32.pte  81.7 MB41xnnpack/efficientnet_v2_s_xnnpack_int8.pte  21.9 MB42```43 44## Compatibility45 46These files are published for the **ExecuTorch v1.4.1** runtime. ExecuTorch47gives no forward compatibility guarantee, so an older runtime may fail to load48them.49 50To use them in React Native ExecuTorch, pass the model constant shipped in the51library's model registry to the corresponding task pipeline. See the52[documentation](https://docs.swmansion.com/react-native-executorch/docs/fundamentals/downloading-models).53 54To load these files in your own ExecuTorch runtime, read the55[compatibility note](https://github.com/pytorch/executorch/blob/main/runtime/COMPATIBILITY.md)56first.57