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

model-migration

Migrate model code and patches to Ascend NPU environments.

31|1|Updated Mar 28, 2026
Search Tags:#migration#training#open-source#patching#ascend-npu#drivingsdk
One-click install
npx skills add https://github.com/Ascend/agent-skills --skill model-migration
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: model-migration
Source: https://github.com/Ascend/agent-skills/tree/main/skills/drivingsdk-ascend-model-migration/model-migration
Command: npx skills add https://github.com/Ascend/agent-skills --skill model-migration

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill helps you migrate open-source model code to Ascend NPU by cloning repositories and applying required patches so training can run on Ascend hardware.

Core Features & Use Cases

  • Clone target model repositories into DrivingSDK/model_examples/<model_name>/ as required.
  • Checkout specific commits and apply NPU patches to align with Ascend runtime.
  • Setup data and weight links to prepare for training.

Quick Start

Clone the target model repository into DrivingSDK/model_examples/<model_name>/, check out the specified commit, apply the patch, and arrange data and weight links before training.

Frequently Asked Questions about model-migration

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

FAQPage Schema
How do I migrate open-source models to Ascend NPU for training?▼

To migrate open-source models to Ascend NPU, clone the target repository into DrivingSDK/model_examples/, checkout the specified commit, apply NPU adaptation patches, and setup data and weight symlinks for training.

When do I need to apply NPU adaptation patches for open-source model training?▼

You need NPU adaptation patches when preparing open-source model repositories to run on Ascend hardware, ensuring the code aligns with the Ascend runtime environment within DrivingSDK before training begins.

Does model migration with DrivingSDK require checking out specific commits?▼

Yes, model migration enforces a defined workflow within DrivingSDK that requires checking out specific repository commits before applying NPU patches to ensure compatibility with the Ascend runtime.

What is the best way to configure data and weight links before training on Ascend NPU?▼

The best way to configure links is to setup data and ckpts symlinks within the cloned DrivingSDK/model_examples/<model_name>/ directory, connecting your datasets and weights directly to the patched model code.

Why does my open-source model fail to train on Ascend NPU after cloning?▼

Training fails if you skip the required NPU adaptation patches. After cloning the repository, you must checkout the specified commit and apply patches so the model aligns with the Ascend runtime environment.

Related Skills

Explore More Software Engineering Skills→
AscendAscend
31

ascend-mmcv-install

Install mmcv-full with NPU support inside Ascend NPU containers.

Official
Intermediate
FeRhodiumFeRhodium
7

architecture-analysis

Analyze PyTorch model architectures for Ascend NPU migration requirements.

Community
Advanced
Lord1EgyptLord1Egypt
2

transformers

Load, run, and fine-tune pre-trained transformer models for NLP, vision, and audio tasks.

Community
Advanced
jackspacejackspace
16

torchdrug

Apply graph neural networks to drug discovery tasks with PyTorch.

Community
Advanced