Software Mansion 发布 whisper-tiny 的 ExecuTorch 多后端导出模型
原标题：software-mansion/whisper-tiny-ExecuTorch
AI 摘要
Software Mansion 在 Hugging Face 发布了 whisper-tiny 的 ExecuTorch 导出版本，提供 CoreML、MLX、Vulkan、XNNPACK 四种后端及 fp16/bf16/int8/fp32 多种精度变体，供 React Native ExecuTorch 库在移动设备上本地运行语音识别。仓库给出 XNNPACK int8 与 fp32 在 LibriSpeech 和 FLEURS 上的 WER/CER 对比，int8 精度基本无损且体积更小。这些文件面向 ExecuTorch v1.4.1 运行时，官方不保证向前兼容。
正文节选
--- license: apache-2.0 pipeline_tag: automatic-speech-recognition library_name: executorch --- # whisper-tiny This repository hosts the **whisper-tiny** 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: [whisper-tiny](https://huggingface.co/openai/whisper-tiny) ## Variants | Path | Backend | Precision | | --- | --- | --- | | `coreml/whisper_tiny_coreml_fp16.