On-device AI inference powered by PyTorch

ExecuTorch is PyTorch's unified solution for deploying AI models on-device—from smartphones to microcontrollers—built for privacy, performance, and portability. It powers Meta's on-device AI across Instagram, WhatsApp, Quest 3, Ray-Ban Meta Smart Glasses, and more.

Deploy LLMs, vision, speech, and multimodal models with the same PyTorch APIs you already know—accelerating research to production with seamless model export, optimization, and deployment. No manual C++ rewrites. No format conversions. No vendor lock-in.

• 🔒 Native PyTorch Export — Direct export from PyTorch. No .onnx, .tflite, or intermediate format conversions. Preserve model semantics.
• ⚡ Production-Proven — Powers billions of users at Meta with real-time on-device inference.
• 💾 Tiny Runtime — 50KB base footprint. Runs on microcontrollers to high-end smartphones.
• 🚀 12+ Hardware Backends — Open-source acceleration for Apple, Qualcomm, ARM, MediaTek, Vulkan, and more.
• 🎯 One Export, Multiple Backends — Switch hardware targets with a single line change. Deploy the same model everywhere.

ExecuTorch uses ahead-of-time (AOT) compilation to prepare PyTorch models for edge deployment:

1. 🧩 Export — Capture your PyTorch model graph with torch.export()
2. ⚙️ Compile — Quantize, optimize, and partition to hardware backends → .pte
3. 🚀 Execute — Load .pte on-device via lightweight C++ runtime

Models use a standardized Core ATen operator set. Partitioners delegate subgraphs to specialized hardware (NPU/GPU) with CPU fallback.

Learn more: How ExecuTorch Works • Architecture Guide

For platform-specific setup (Android, iOS, embedded systems), see the Quick Start documentation for additional info.

import torch
from executorch.exir import to_edge_transform_and_lower
from executorch.backends.xnnpack.partition.xnnpack_partitioner import XnnpackPartitioner

# 1. Export your PyTorch model
model = MyModel().eval()
example_inputs = (torch.randn(1, 3, 224, 224),)
exported_program = torch.export.export(model, example_inputs)

# 2. Optimize for target hardware (switch backends with one line)
program = to_edge_transform_and_lower(
    exported_program,
    partitioner=[XnnpackPartitioner()]  # CPU | CoreMLPartitioner() for iOS | QnnPartitioner() for Qualcomm
).to_executorch()

# 3. Save for deployment
with open("model.pte", "wb") as f:
    f.write(program.buffer)

# Test locally via ExecuTorch runtime's pybind API (optional)
from executorch.runtime import Runtime
runtime = Runtime.get()
method = runtime.load_program("model.pte").load_method("forward")
outputs = method.execute([torch.randn(1, 3, 224, 224)])

C++

#include <executorch/extension/module/module.h>
#include <executorch/extension/tensor/tensor.h>

Module module("model.pte");
auto tensor = make_tensor_ptr({2, 2}, {1.0f, 2.0f, 3.0f, 4.0f});
auto outputs = module.forward(tensor);

Swift (iOS)

import ExecuTorch

let module = Module(filePath: "model.pte")
let input = Tensor<Float>([1.0, 2.0, 3.0, 4.0], shape: [2, 2])
let outputs = try module.forward(input)

Kotlin (Android)

val module = Module.load("model.pte")
val inputTensor = Tensor.fromBlob(floatArrayOf(1.0f, 2.0f, 3.0f, 4.0f), longArrayOf(2, 2))
val outputs = module.forward(EValue.from(inputTensor))

Export Llama models using the export_llm script or Optimum-ExecuTorch:

# Using export_llm
python -m executorch.extension.llm.export.export_llm --model llama3_2 --output llama.pte

# Using Optimum-ExecuTorch
optimum-cli export executorch \
  --model meta-llama/Llama-3.2-1B \
  --task text-generation \
  --recipe xnnpack \
  --output_dir llama_model

Run on-device with the LLM runner API:

C++

#include <executorch/extension/llm/runner/text_llm_runner.h>

auto runner = create_llama_runner("llama.pte", "tiktoken.bin");
executorch::extension::llm::GenerationConfig config{
    .seq_len = 128, .temperature = 0.8f};
runner->generate("Hello, how are you?", config);

Swift (iOS)

import ExecuTorchLLM

let runner = TextRunner(modelPath: "llama.pte", tokenizerPath: "tiktoken.bin")
try runner.generate("Hello, how are you?", Config {
    $0.sequenceLength = 128
}) { token in
    print(token, terminator: "")
}

Kotlin (Android) — API Docs • Demo App

val llmModule = LlmModule("llama.pte", "tiktoken.bin", 0.8f)
llmModule.load()
llmModule.generate("Hello, how are you?", 128, object : LlmCallback {
    override fun onResult(result: String) { print(result) }
    override fun onStats(stats: String) { }
})

For multimodal models (vision, audio), use the MultiModal runner API which extends the LLM runner to handle image and audio inputs alongside text. See Llava and Voxtral examples.

See examples/models/llama for complete workflow including quantization, mobile deployment, and advanced options.

Next Steps:

See Backend Documentation for detailed hardware requirements and optimization guides. For desktop/laptop GPU inference with CUDA and Metal, see the Desktop Guide. For Zephyr RTOS integration, see the Zephyr Guide.

ExecuTorch powers on-device AI at scale across Meta's family of apps, VR/AR devices, and partner deployments. View success stories →

LLMs: Llama 3.2/3.1/3, Qwen 3, Phi-4-mini, LiquidAI LFM2

Multimodal: Llava (vision-language), Voxtral (audio-language), Gemma (vision-language)

Vision/Speech: MobileNetV2, DeepLabV3, Whisper

Resources: examples/ directory • executorch-examples out-of-tree demos • Optimum-ExecuTorch for HuggingFace models • Unsloth for fine-tuned LLM deployment

ExecuTorch provides advanced capabilities for production deployment:

• Quantization — Built-in support via torchao for 8-bit, 4-bit, and dynamic quantization
• Memory Planning — Optimize memory usage with ahead-of-time allocation strategies
• Developer Tools — ETDump profiler, ETRecord inspector, and model debugger
• Selective Build — Strip unused operators to minimize binary size
• Custom Operators — Extend with domain-specific kernels
• Dynamic Shapes — Support variable input sizes with bounded ranges

See Advanced Topics for quantization techniques, custom backends, and compiler passes.

We welcome contributions from the community!

ExecuTorch is BSD licensed, as found in the LICENSE file.