skills.rest
Ascend
AscendOFFICIAL
Explore all skills by this author →
WebsiteGitHub

atc-model-converter

A structured approach to managing your finances and hitting your savings targets through automated transfers.

31|1|Updated Mar 28, 2026
Search Tags:#end-to-end#onnx#model-conversion#ascend#cann#atc
One-click install
npx skills add https://github.com/Ascend/agent-skills --skill atc-model-converter-ascend
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: atc-model-converter
Source: https://github.com/Ascend/agent-skills/tree/main/skills/atc-model-converter
Command: npx skills add https://github.com/Ascend/agent-skills --skill atc-model-converter-ascend

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires numpy, onnxruntime, ais_bench, torch, torchvision, onnx, and includes scripts (resource) and references (resource) components.

What problem does it solve?

This Skill provides a complete ATC-based workflow to convert PyTorch/ONNX models into Ascend OM, enabling deployment and end-to-end inference on Ascend devices.

Core Features & Use Cases

  • Workflow discovery: detects input shapes, exports ONNX, performs ATC conversion, and generates an end-to-end inference script.
  • Multi-CANN version support: handles 8.1.RC1, 8.3.RC1, and 8.5.0+ environments for broad compatibility.
  • End-to-end automation: produces reusable end-to-end inference scripts and a reproducible README for quick replication.

Quick Start

Provide a ready-to-run end-to-end ATC workflow starting from a PyTorch/ONNX model through ONNX export, ATC conversion, and end-to-end inference script generation.

Frequently Asked Questions about atc-model-converter

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

FAQPage Schema
How do I convert a PyTorch model to Ascend OM format?▼

This Skill converts PyTorch models to Ascend OM by first exporting the model to ONNX and then using ATC for conversion, yielding a production-ready OM file for deployment on Ascend devices.

What is the end-to-end ATC model conversion workflow?▼

The end-to-end ATC model conversion workflow detects input shapes, exports the model to ONNX, performs ATC conversion to OM, and generates a reusable end-to-end inference script for Ascend hardware.

Does ATC model conversion support different CANN versions?▼

Yes, this ATC model conversion process supports CANN versions 8.1.RC1, 8.3.RC1, and 8.5.0+ environments, ensuring broad compatibility and robust error recovery during deployment.

Can I use ONNX models directly for Ascend OM conversion?▼

Yes, you can use ONNX models directly for Ascend OM conversion. The Skill accepts ONNX inputs, validates the environment, and uses ATC to generate the final OM model for inference.

How do I generate an inference script for my converted OM model?▼

You generate an inference script for your converted OM model using the end-to-end automation feature, which produces reusable inference scripts and a reproducible README alongside the ATC conversion output.

Why does ONNX export fail during Ascend model conversion?▼

ONNX export may fail during Ascend model conversion due to environment issues. This Skill ensures correct CANN version handling and environment validation to provide robust error recovery for ONNX export.

Related Skills

Explore More Software Engineering Skills→
AscendAscend
31

cann-operator-env-config

Install and configure the CANN toolkit on Ascend NPU hardware.

Official
Advanced
ascend-ai-codingascend-ai-coding
156

atc-model-converter

Convert ONNX models to Ascend OM format with ATC and validate via AIS Bench.

Official
Advanced
4lerman4lerman

senior-computer-vision

Design, train, and deploy computer vision models with PyTorch and ONNX.

Community
Advanced
juliuss1907juliuss1907

llama-cpp

Run local GGUF inference with llama.cpp models from Hugging Face Hub.

Community
Advanced