Find Your Path in ExecuTorch

ExecuTorch serves a wide range of users — from ML engineers taking their first steps in on-device inference, to embedded systems developers targeting bare-metal microcontrollers, to researchers pushing the boundaries of LLM deployment. This page helps you navigate directly to the content most relevant to your experience level, goals, and target platform.

---

Step 1: What best describes your experience?

🟢 New to ExecuTorch

Beginner

You are familiar with PyTorch but have not yet deployed a model to an edge device. You want a clear, guided path from installation to your first on-device inference.

→ Beginner Pathway

Beginner Pathway

🟡 I know the basics

Quick Start

You have some experience with model export or mobile ML, and want to get a model running as fast as possible without reading through every concept first.

→ Quick Start Pathway

Quick Start Pathway

🔴 Experienced / Production

Advanced

You are building production systems, integrating custom backends, optimizing for constrained hardware, or working with LLMs on edge devices.

→ Advanced Pathway

Advanced Pathway

---

Step 2: What is your primary goal?

Use the decision matrix below to jump directly to the most relevant section based on your goal and target platform.

ExecuTorch Decision Matrix

Goal	Android	iOS / macOS	Desktop / Server	Embedded / MCU
Run a pre-exported model quickly	Android	iOS	Getting Started with ExecuTorch	Embedded Systems
Export my own PyTorch model	Model Export and Lowering	Model Export and Lowering	Getting Started with ExecuTorch	Model Export and Lowering
Deploy an LLM (Llama, Phi, etc.)	Llama on ExecuTorch	Running LLMs on iOS	LLMs	—
Use hardware acceleration (NPU/GPU)	Qualcomm AI Engine Backend	Core ML Backend	Backends	Backends
Integrate a custom backend delegate	Backend Development	Backend Development	Backend Development	Backend Development
Profile and debug my model	Developer Tools Usage Tutorials	Developer Tools Usage Tutorials	Developer Tools Usage Tutorials	Developer Tools Usage Tutorials
Build from source	Building from Source	Building from Source	Building from Source	Building from Source

---

Step 3: What is your role?

Different roles have different entry points into ExecuTorch. Select the one that best matches your background.

🤖 ML Engineer

You work primarily in Python, train models with PyTorch, and want to deploy them efficiently to edge devices.

Start here:

• Getting Started with ExecuTorch — Installation and first export

• Model Export and Lowering — Model export and lowering

• Backends — Choose the right backend

• Quantization — Reduce model size and latency

📱 Mobile Developer

You build Android or iOS applications and need to integrate an on-device ML model into your app.

Start here:

• Android — Android integration guide

• iOS — iOS integration guide

• Getting Started with ExecuTorch — Export a model for your platform

• Using ExecuTorch on Android — Full Android API reference

⚙️ Backend / Systems Developer

You are implementing a hardware backend, writing C++ runtime integrations, or contributing to ExecuTorch internals.

Start here:

• Backend Development — Backend delegate development

• Integrating a Backend Delegate into ExecuTorch — Integrating a backend

• Detailed C++ Runtime APIs Tutorial — C++ runtime APIs

• New Contributor Guide — Contributing to ExecuTorch

🔌 Embedded Developer

You target microcontrollers, DSPs, or other resource-constrained hardware where memory and compute are tightly limited.

Start here:

• Embedded Systems — Embedded platforms overview

• Backends — Available embedded backends

• Portable C++ Programming — Portable C++ for constrained devices

• Pico2: A simple MNIST Tutorial — MNIST on Raspberry Pi Pico2

---

Step 4: What is your model’s status?

The right workflow depends on whether you are starting from scratch, using a supported model, or working with a custom architecture.

Model Status Routing

Model Status	Recommended Path
Using a supported LLM (Llama, Phi, Qwen, SmolLM)	Use the Exporting LLMs script for a streamlined export with quantization and optimization built in. Pre-exported models are also available on HuggingFace ExecuTorch Community <https://huggingface.co/executorch-community>_.
Using a HuggingFace model	Use Exporting LLMs with HuggingFace’s Optimum ExecuTorch (Optimum ExecuTorch) for broad HuggingFace model support with familiar APIs.
Using a custom PyTorch model	Follow Getting Started with ExecuTorch for the standard export flow, then consult Model Export and Lowering for advanced lowering options.
Model requires dynamic shapes	See the dynamic shapes section in Model Export and Lowering and the Exporting to ExecuTorch for constraints.
Model uses unsupported operators	Consult Kernel Registration to register custom kernels, or Custom Compiler Passes and Partitioners for graph-level transformations.
Pre-exported .pte file available	Skip export entirely and go directly to Getting Started with ExecuTorch (Running on Device section) or your platform guide.

---

Not sure where to start?

If you are completely new to ExecuTorch, the recommended entry point is the Getting Started with ExecuTorch guide, which walks through installation, exporting a MobileNet V2 model, and running inference in under 15 minutes.

For a high-level conceptual overview before diving into code, read ExecuTorch Overview and How ExecuTorch Works.