export

Convert PyTorch models into portable .pte files for ExecuTorch deployments.

Updated Mar 25, 2026
One-click install
npx skills add https://github.com/huahai2022/execu-xnnpack --skill export-huahai2022
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: export
Source: https://github.com/huahai2022/execu-xnnpack/tree/main/.claude/skills/export
Command: npx skills add https://github.com/huahai2022/execu-xnnpack --skill export-huahai2022

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

The Skill enables converting PyTorch models into a compact, edge-friendly .pte format for ExecuTorch deployments, simplifying model export and deployment pipelines.

Core Features & Use Cases

  • Export PyTorch models to the .pte format for edge deployment.
  • Provide guidance for model-specific export workflows and debugging.
  • Real-world scenario: convert a trained Llama or Whisper model to .pte for on-device inference.

Quick Start

Run an end-to-end export of a sample model to .pte and verify the generated file.

Frequently Asked Questions about export

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I export a PyTorch model to .pte format for edge deployment?

To export a PyTorch model to .pte format, you use the export, to_edge_transform_and_lower functions to lower the model for ExecuTorch, then write the resulting .pte file to disk for on-device inference.

Can I convert Llama or Whisper models to .pte format for on-device inference?

Yes, you can convert Llama or Whisper models to .pte format. The export process supports lowering these PyTorch models into a portable edge format for ExecuTorch deployments and offline inference.

What is the .pte format used for in ExecuTorch?

The .pte format is a compact, edge-friendly file format used for ExecuTorch deployments. It enables on-device inference by providing a portable representation of lowered PyTorch models on edge devices.

Do I need to understand to_edge_transform_and_lower to export PyTorch models?

Yes, understanding to_edge_transform_and_lower is required. This function is a core part of the workflow to lower PyTorch models into the portable .pte edge format during the export pipeline.

Does this export process support offline inference on edge devices?

Yes, the export process supports offline inference. By converting trained PyTorch models into the compact .pte format, the workflow enables deployment pipelines for edge devices without requiring continuous network connectivity.

Why is my PyTorch model export to .pte format failing?

Export to .pte format can fail if the model-specific export workflow is incorrect. Proper debugging and understanding of the export and to_edge_transform_and_lower steps are required to successfully write the .pte file.