#6767th of 206 in ML frameworks and libraries
executorch
PyTorch runtime for running exported models on phones, browsers and microcontrollers
- Stars
- 5.2k
- License
- custom license
- Last commit
- Nov 2026
- Last release
- Sep 2026
- Language
- Python
4,481 on 22 Apr 20265,173 on 11 Nov 2026
From GitHub's daily star counts.
Checked on GitHub.
Overview
ExecuTorch takes a PyTorch model captured with torch.export, lowers it to a backend-specific .pte file, and runs it through C++, Python, Swift/Objective-C, Kotlin/Java or JavaScript APIs. Backends include XNNPACK, Core ML, Vulkan, Qualcomm, MediaTek, OpenVINO, Arm Cortex-M and Ethos-U. It also ships LLM, speech and vision examples and an experimental OpenAI-compatible local server.
Who it is for: Engineers deploying PyTorch models to mobile, embedded and desktop targets
Strengths
- Starts from torch.export, so existing PyTorch models need no separate framework
- Backends span XNNPACK, Core ML, Vulkan, Qualcomm, MediaTek, OpenVINO, Arm Ethos-U, Cadence
- Selective build links only the operators, kernels and delegates a deployment needs
- Stable-API .pte files are guaranteed to load for at least one following non-patch release
Weaknesses
- Each hardware target needs its own backend-specific .pte export
- Deployment depends on torch.export capture and operator coverage; some models need decompositions
- CUDA, Metal, MLX, WebGPU and WASM paths are labeled experimental
- Features described on main/nightly may change before release
What it needs
- GPU optional
- Needs PyTorch (torch.export), Python 3.11-3.15
- Models: PyTorch models via torch.export (.pte), LLMs, speech models, vision models, Hugging Face models via Optimum ExecuTorch
What changed
From the release notes. No flag means the note does not say so, not that the upgrade is safe.
Also in ML frameworks and libraries
See all 206| Rank | Project | Score |
|---|---|---|
| 1 | transformersPython library defining open models for text, vision, audio and multimodal tasks | 94 out of 100 |
| 2 | ultralyticsPython library and CLI for training and running YOLO vision models | 94 out of 100 |
| 3 | pytorchPython tensor library with GPU support and tape-based autograd | 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 |