.TFLITE 文件扩展名
TensorFlow Lite Model
| 开发者 | TensorFlow |
| 受欢迎程度 |
3.3 | 4 票 |
什么是 TFLITE 文件？
A TFLITE file is a machine learning model used by TensorFlow Lite, an app that runs machine learning models on mobile devices, microcontrollers, and embedded systems. The model a TFLITE file contains may be used for image classification, object detection, text classification, or another type of machine learning task. TensorFlow Lite users create TFLITE files with TensorFlow Lite Model Maker or TensorFlow Lite Converter.
更多信息
TensorFlow Lite is a lightweight version of TensorFlow's machine learning platform, designed for use on mobile, embedded, and edge devices. While both TensorFlow Lite and TensorFlow use machine learning models, the models they use are saved in different formats. TensorFlow Lite models are saved in the TFLITE format, which includes optimizations that reduce a model's size with little to no effect on accuracy.
To create a TFLITE file, developers can use:
- TensorFlow Lite Converter to convert an existing TensorFlow model to the TFLITE format (some models cannot be converted)
- TensorFlow Lite Model Maker to create a TFLITE model from scratch
Several example TFLITE files are available for download from TensorFlow's website and GitHub page.
注: By default, TFLITE files created using TensorFlow Model Maker contain metadata, while TFLITE files created using TensorFlow Lite Converter do not.
如何打开 TFLITE 文件
TFLITE files are not meant to be opened. For instructions on how to use TFLITE files with TensorFlow Lite (Android, iOS, Linux), refer to the app's developer documentation.