Atb Nnal Installer logo

Atb Nnal Installer

Organization
Ascend
atb-nnal-installer

昇腾 NPU NNAL(ATB 加速库)安装技能。依赖 cann-operator-env-config 提供 Toolkit+Kernels 环境,本技能仅负责 NNAL 包的安装、环境变量配置与验证。

Overview

PublisherAscend
Repositoryagent-skills
Skill nameatb-nnal-installer
Stars
54
Forks
4
Bundled files
Instructions only
LicenseMulanPSL-2.0
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • Self-contained

    Everything the model needs lives in the instructions — no extra files to sync.

  • Open source

    Published by Ascend on GitHub. Read the source before you install it.

Installation

Install the Atb Nnal Installer AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/Ascend/agent-skills/tree/master/skills/ascend-transformer-boost/skills/atb-nnal-installer
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/Ascend/agent-skills.git /tmp/agent-skills
mkdir -p .claude/skills
cp -r /tmp/agent-skills/skills/ascend-transformer-boost/skills/atb-nnal-installer .claude/skills/atb-nnal-installer
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Atb Nnal Installer in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Atb Nnal Installer on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Atb Nnal Installer is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

CANN NNAL 安装部署

功能概述

在昇腾 NPU 环境中安装 NNAL(ATB 加速库)。本技能假设 Toolkit+Kernels 已由 cann-operator-env-config 技能安装完毕。

调用时机

  • 用户已完成 Toolkit+Kernels 安装(通过 cann-operator-env-config),需要安装 NNAL
  • 用户在 Docker 容器中需要安装 NNAL
  • 用户需要安装或升级特定版本的 NNAL

前置条件

必须已执行 cann-operator-env-config 技能,确保:

  • Toolkit 已安装,/usr/local/Ascend/ascend-toolkit/set_env.sh 存在
  • Kernels(ops)已安装

⚠️ 路径约束(必须执行)

执行此技能前,必须从用户处获取以下路径:

  • <INSTALL_PATH>: NNAL 安装目标路径(如 /usr/local/Ascend)

若用户未提供 INSTALL_PATH,立即停止并打断,要求用户提供安装路径。

脚本校验示例:

bash
if [ -z "$INSTALL_PATH" ]; then
    echo "ERROR: INSTALL_PATH 未设置,请提供 NNAL 安装路径。"
    exit 1
fi

详细内容

第1步:检查 NPU 驱动

bash
npu-smi info

若 npu-smi 不可用或未检测到 NPU 设备,停止安装并报错。

第2步:调用 cann-operator-env-config(如尚未安装 Toolkit+Kernels)

如果 Toolkit+Kernels 尚未安装,首先调用 cann-operator-env-config 技能安装 Toolkit 和 Kernels:

  • 该技能位于 /home/s30073260/swiglu_quant/agent-skills-ivan-fork/skills/cann-operator-env-config/
  • 提供 3 种安装方式:离线 run 包、conda 在线、yum 在线
  • 安装完成后,Toolkit 和 Kernels 将位于 /usr/local/Ascend/

第3步:Source Toolkit 环境变量

bash
source /usr/local/Ascend/ascend-toolkit/set_env.sh

关键:NNAL 安装脚本依赖 ASCEND_TOOLKIT_HOME 等环境变量,必须先 source。

验证环境变量:

bash
echo $ASCEND_TOOLKIT_HOME
# 应输出: /usr/local/Ascend/ascend-toolkit/latest 或自定义路径

第4步:安装 NNAL

注意:以下命令使用用户提供的 <INSTALL_PATH>,如未提供则已在路径约束步骤中打断。

方式一:run 包安装(推荐)
bash
chmod +x Ascend-cann-nnal_*.run
./Ascend-cann-nnal_*.run --install --install-path="${INSTALL_PATH}"
方式二:从 Docker 镜像提取
bash
CANN_IMAGE="quay.io/ascend/cann:<版本>-<芯片>-<OS>-py<Python版本>"

# 示例
# CANN_IMAGE="quay.io/ascend/cann:9.0.0-beta.2-910-openeuler24.03-py3.11"

docker create --name cann_nnal_temp "${CANN_IMAGE}"
docker cp cann_nnal_temp:/usr/local/Ascend/nnal "${INSTALL_PATH}/"
docker rm cann_nnal_temp

第5步:加载环境变量

bash
source "${INSTALL_PATH}/nnal/atb/set_env.sh"

持久化到 ~/.bashrc:

bash
cat >> ~/.bashrc << EOF
source ${INSTALL_PATH}/nnal/atb/set_env.sh
EOF

第6步:安装验证

bash
# 检查 NNAL 目录结构
ls -lh ${INSTALL_PATH}/nnal/atb/

# 验证 Toolkit 版本
cat /usr/local/Ascend/ascend-toolkit/latest/version.cfg 2>/dev/null || \
    cat /usr/local/Ascend/ascend-toolkit/latest/arm64-linux/ascend_toolkit_install.info

# 验证 NNAL 版本
cat ${INSTALL_PATH}/nnal/atb/version 2>/dev/null || echo "检查 NNAL 安装目录"

# 验证关键命令
which bisheng && bisheng --version
which atc && atc --version

# 验证 NPU 状态
npu-smi info

环境变量汇总

安装完成后,以下环境变量应正确设置:

变量值来源
ASCEND_TOOLKIT_HOME/usr/local/Ascend/ascend-toolkit/latestset_env.sh (Toolkit)
ATB_HOME_PATH${INSTALL_PATH}/nnal/atbset_env.sh (NNAL)
LD_LIBRARY_PATH含 ${INSTALL_PATH}/nnal/atb/libset_env.sh (NNAL)

参考信息

项目参考值
NNAL run 包大小~500-600MB
安装耗时约 5-10 分钟
总计(含 Toolkit+Kernels)约 15-20 分钟(3 包 ~3.6GB)

部分 run 包会询问确认,使用 yes | 管道实现非交互安装:

bash
yes | ./Ascend-cann-nnal_*.run --install --install-path="${INSTALL_PATH}"

安装时若出现 "driver package is not installed" 警告,属于正常现象——Docker 容器使用宿主机驱动,无需安装驱动包。


常见问题

原因:NPU 驱动未安装或 Docker 容器未挂载 NPU 设备。

解决:

  1. 确认宿主机 NPU 驱动正常:宿主机执行 npu-smi info
  2. Docker 容器启动时添加 --device=/dev/davinciX 参数

原因:Toolkit 未安装或 set_env.sh 路径不正确。

解决:

  1. 确认 /usr/local/Ascend/ascend-toolkit/latest/set_env.sh 存在
  2. 如不存在,返回执行 cann-operator-env-config 安装 Toolkit+Kernels

原因:安装路径需 root 权限。

解决:

bash
# 使用 root 权限安装
sudo ./Ascend-cann-nnal_*.run --install --install-path=/usr/local/Ascend

执行结果

⚠️ 执行后填写:技能执行完成后,参照下方格式填写实际执行结果。

检查点检查表

步骤检查点描述状态
1NPU 驱动正常 (npu-smi info)✅ PASS / ❌ FAIL / ⏭️ SKIP
2Toolkit+Kernels 已安装(通过 cann-operator-env-config)✅ PASS / ❌ FAIL / ⏭️ SKIP
3ASCEND_TOOLKIT_HOME 已设置✅ PASS / ❌ FAIL / ⏭️ SKIP
4NNAL 安装成功✅ PASS / ❌ FAIL / ⏭️ SKIP
5环境变量已加载✅ PASS / ❌ FAIL / ⏭️ SKIP
6NNAL 版本验证通过✅ PASS / ❌ FAIL / ⏭️ SKIP
6bisheng / atc 可用✅ PASS / ❌ FAIL / ⏭️ SKIP
6npu-smi info 正常✅ PASS / ❌ FAIL / ⏭️ SKIP

VERDICT: ✅ SUCCESS / ⚠️ PARTIAL / ❌ FAILED / ⏭️ SKIPPED

问题列表(若有)

等级检查点问题描述建议
🔴 CRITICAL---
🟡 WARNING---

执行摘要

  • 执行时间:
  • 执行环境:
  • NNAL 版本:
  • 持续时间:
  • 通过率:

Frequently asked questions

What does the Atb Nnal Installer AI skill do?

昇腾 NPU NNAL(ATB 加速库)安装技能。依赖 cann-operator-env-config 提供 Toolkit+Kernels 环境,本技能仅负责 NNAL 包的安装、环境变量配置与验证。

Why use Atb Nnal Installer on TypingMind?

Because you install it once and use it with any model. Atb Nnal Installer is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Atb Nnal Installer in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/Ascend/agent-skills/tree/master/skills/ascend-transformer-boost/skills/atb-nnal-installer. TypingMind reads its SKILL.md and installs it as a skill you can enable per chat.

Which AI models can use Atb Nnal Installer?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Atb Nnal Installer?

As many as you like. As long as a model supports skills, you can use Atb Nnal Installer with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Atb Nnal Installer AI skill free?

Yes. It is published on GitHub by Ascend under the MulanPSL-2.0 license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇