React Native ExecuTorch 发布 gemma-4 端侧模型
原标题：software-mansion/react-native-executorch-gemma-4
AI 摘要
Software Mansion 在 Hugging Face 发布了 react-native-executorch-gemma-4 仓库，将 Google DeepMind 的 gemma-4（基于 gemma-4-E2B）导出为 ExecuTorch .pte 程序，供 React Native ExecuTorch 库在设备端运行。仓库提供 mlx（int4，4096 上下文）、vulkan（8da4w，2048）和 xnnpack（8da4w，4096）三种后端变体，文件面向 ExecuTorch v1.4.1 运行时，不保证向前兼容。
正文节选
--- license: gemma pipeline_tag: text-generation tags: - google - deepmind - gemma4 - edge base_model: - google/gemma-4-E2B library_name: executorch --- # gemma-4 This repository hosts the **gemma-4** 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. ## Variants | Path | Backend | Precision | Context length | | --- | --- | --- | --- | | `e2b/mlx/gemma_4_e2b_mlx_int4.pte` | ml