Apply the model card standard

README.md

@@ -6,27 +6,59 @@ tags:
 library_name: executorch
 ---
 
-# Introduction
+# qwen-3
 
-This repository hosts the **Qwen 3** models for the [React Native ExecuTorch](https://www.npmjs.com/package/react-native-executorch) library. It includes **unquantized** and **quantized** versions of the Qwen model in `.pte` format, ready for use in the **ExecuTorch** runtime.
+This repository hosts the **qwen-3** 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.
 
-If you'd like to run these models in your own ExecuTorch runtime, refer to the [official documentation](https://pytorch.org/executorch/stable/index.html) for setup instructions.
+## Variants
 
-## Compatibility
-
-If you intend to use this model outside of React Native ExecuTorch, make sure your runtime is compatible with the **ExecuTorch** version used to export the `.pte` files. For more details, see the compatibility note in the [ExecuTorch GitHub repository](https://github.com/pytorch/executorch/blob/11d1742fdeddcf05bc30a6cfac321d2a2e3b6768/runtime/COMPATIBILITY.md?plain=1#L4). If you work with React Native ExecuTorch, the constants from the library will  guarantee compatibility with runtime used behind the scenes.
+| Path | Backend | Precision |
+| --- | --- | --- |
+| `0_6b/mlx/qwen_3_0_6b_mlx_int4.pte` | mlx | 4w |
+| `0_6b/xnnpack/qwen_3_0_6b_xnnpack_bf16.pte` | xnnpack | bf16 |
+| `0_6b/xnnpack/qwen_3_0_6b_xnnpack_8da4w.pte` | xnnpack | 8da4w |
+| `1_7b/mlx/qwen_3_1_7b_mlx_int4.pte` | mlx | 4w |
+| `1_7b/xnnpack/qwen_3_1_7b_xnnpack_bf16.pte` | xnnpack | bf16 |
+| `1_7b/xnnpack/qwen_3_1_7b_xnnpack_8da4w.pte` | xnnpack | 8da4w |
+| `4b/mlx/qwen_3_4b_mlx_int4.pte` | mlx | 4w |
+| `4b/xnnpack/qwen_3_4b_xnnpack_bf16.pte` | xnnpack | bf16 |
+| `4b/xnnpack/qwen_3_4b_xnnpack_8da4w.pte` | xnnpack | 8da4w |
 
-These models were exported using `v0.6.0` version and **no forward compatibility** is guaranteed. Older versions of the runtime may not work with these files.
+## Repository structure
 
-### Repository Structure
+```
+0_6b/mlx/config.json                        1.3 kB
+0_6b/mlx/qwen_3_0_6b_mlx_int4.pte           599 MB
+0_6b/xnnpack/config.json                    2.3 kB
+0_6b/xnnpack/qwen_3_0_6b_xnnpack_8da4w.pte  482 MB
+0_6b/xnnpack/qwen_3_0_6b_xnnpack_bf16.pte   1.1 GB
+1_7b/mlx/config.json                        1.3 kB
+1_7b/mlx/qwen_3_1_7b_mlx_int4.pte           1.4 GB
+1_7b/xnnpack/config.json                    2.3 kB
+1_7b/xnnpack/qwen_3_1_7b_xnnpack_8da4w.pte  1.2 GB
+1_7b/xnnpack/qwen_3_1_7b_xnnpack_bf16.pte   3.2 GB
+4b/mlx/config.json                          1.3 kB
+4b/mlx/qwen_3_4b_mlx_int4.pte               2.7 GB
+4b/xnnpack/config.json                      2.3 kB
+4b/xnnpack/qwen_3_4b_xnnpack_8da4w.pte      2.5 GB
+4b/xnnpack/qwen_3_4b_xnnpack_bf16.pte       7.5 GB
+config.json                                 28 B
+tokenizer.json                              10.9 MB
+tokenizer_config.json                       7.8 kB
+```
 
-The repository is organized into three main directories:
+## Compatibility
 
-- `qwen-3-0.6B`
-- `qwen-3-1.7B`
-- `qwen-3-4B`
+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.
 
-Each directory contains different versions of the model, including **quantized**, and the **original** models.
+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/loading-models).
 
-- The `.pte` file should be passed to the `modelSource` parameter.
-- The tokenizer for the models is available within the repo root, under `tokenizer.json` and `tokenizer_config.json`
+To load these files in your own ExecuTorch runtime, read the
+[compatibility note](https://github.com/pytorch/executorch/blob/main/runtime/COMPATIBILITY.md)
+first.