#132132nd of 168 in ML frameworks and libraries
react-native-fast-tflite
TensorFlow Lite inference for React Native with zero-copy buffers
- Stars
- 1.2k
- License
- MIT
- Last commit
- Sep 2026
- Last release
- Apr 2026
- Language
- TypeScript
1,161 on 21 Apr 20261,243 on 10 Nov 2026
From GitHub's daily star counts.
Checked on GitHub.
Overview
A React Native library that runs .tflite models on-device through the TensorFlow Lite C/C++ core API, built on Nitro Modules. Inputs and outputs are passed as ArrayBuffers without copying, models can be loaded from the bundle, a file or a URL, and CoreML, Android GPU and NNAPI delegates are optional. It integrates with VisionCamera frame processors for real-time camera inference.
Who it is for: React Native developers running TFLite models on-device
Strengths
- Zero-copy ArrayBuffer input and output via the low-level TFLite C/C++ API
- Models load from bundle, file:// or http(s) URL and can be swapped at runtime
- Optional CoreML, Android GPU and NNAPI delegates, with an Expo config plugin
- Synchronous runSync works inside VisionCamera frame processors
Weaknesses
- You must interpret raw tensor bytes yourself; shapes are inspected manually, e.g. in Netron
- Camera frames need a separate resizer package to match the model input size
- CoreML delegate does not support all model operations
- NNAPI is deprecated on Android 15; GPU libraries may need manifest entries
What it needs
- GPU optional
- Needs react-native-nitro-modules, react-native-vision-camera (optional), react-native-vision-camera-resizer (optional)
- Models: TensorFlow Lite (.tflite)
Quick answers
Is react-native-fast-tflite open source?
Yes. Its code is on GitHub under the MIT license.
Does react-native-fast-tflite need a GPU?
No. It runs on a CPU and can use a GPU when you have one.
Also in ML frameworks and libraries
See all 168| Rank | Project | Score |
|---|---|---|
| 1 | transformersPython library defining open models for text, vision, audio and multimodal tasks | 94 out of 100 |
| 2 | pytorchPython tensor library with GPU support and tape-based autograd | 93 out of 100 |
| 3 | ultralyticsPython library and CLI for training and running YOLO vision models | 93 out of 100 |
| #4 | pytorch-image-modelsPyTorch library of image models, pretrained weights and training scripts | 93 out of 100 |
| #5 | supervisionPython toolkit for loading, annotating and counting computer vision detections | 92 out of 100 |