[CI] replace deprecated pynvml with nvidia-ml-py

[ooooo-create]

Motivation

https://github.com/gpuopenanalytics/pynvml has been marked as deprecated by the official repository, and now relies on nvidia-ml-py.

Modifications

• Replace pynvml with nvidia-ml-py in all requirements files
• Update documentation to reflect the new package name

nvidia-ml-py 通过 import pynvml 进行使用，所以 import 部分不需要进行修改
nvidia 的匹配已经包含了 nvidia-ml-py，故没有更新 fastdeploy/collect_env.py，同时兼容性保留 pynvml 作为依赖展示，判断 nvidia-ml-py 是否由 pynvml 引入

Usage or Command

No new commands added. FastDeploy installation will pull nvidia-ml-py from pip instead of pynvml.

Accuracy Tests

N/A - this is a dependency replacement for CI/CD environment only. No core model inferences varied.

Checklist

• Add at least a tag in the PR title.
• Format your code, run pre-commit before commit.
• Add unit tests. Please write the reason in this PR if no unit tests.
• Provide accuracy results.
• If the current PR is submitting to the release branch, make sure the PR has been submitted to the develop branch, then cherry-pick it to the release branch with the [Cherry-Pick] PR tag.

[paddle-bot]

Thanks for your contribution!

[Copilot]

Pull request overview

该 PR 旨在将已被标记为 deprecated 的 pynvml 依赖替换为其官方迁移目标 nvidia-ml-py，并同步更新环境收集脚本与相关文档示例，以便在安装与 collect_env 输出中反映新的发行包名。

Changes:

• 在多份 requirements 文件中将依赖从 pynvml 替换为 nvidia-ml-py
• 更新 fastdeploy/collect_env.py 的 pip/conda 过滤模式以匹配 nvidia-ml-py
• 更新中英文 collect-env 文档中的示例输出条目

Reviewed changes

Copilot reviewed 6 out of 7 changed files in this pull request and generated 4 comments.

Show a summary per file

File	Description
requirements.txt	主依赖列表将 pynvml 替换为 nvidia-ml-py
requirements_metaxgpu.txt	metaxgpu 环境依赖同步替换
requirements_iluvatar.txt	iluvatar 环境依赖同步替换
requirements_dcu.txt	dcu 环境依赖同步替换
fastdeploy/collect_env.py	collect-env 的 pip/conda 包匹配模式更新为 nvidia-ml-py
docs/cli/collect-env.md	英文文档示例输出更新
docs/zh/cli/collect-env.md	中文文档示例输出更新

[ooooo-create]

PR 描述模板已经补全，并且推送了一个空 commit 来重新触发流水线。

[codecov-commenter]

Codecov Report

✅ All modified and coverable lines are covered by tests.
⚠️ Please upload report for BASE (develop@dae246e). Learn more about missing BASE report.

Additional details and impacted files

@@            Coverage Diff             @@
##             develop    #7290   +/-   ##
==========================================
  Coverage           ?   71.70%           
==========================================
  Files              ?      419           
  Lines              ?    57857           
  Branches           ?     9077           
==========================================
  Hits               ?    41484           
  Misses             ?    13543           
  Partials           ?     2830           

Flag	Coverage Δ
GPU	71.70% <ø> (?)

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[Copilot]

Pull request overview

Copilot reviewed 6 out of 7 changed files in this pull request and generated 2 comments.

[ooooo-create]

/re-run all-failed

[Copilot]

Pull request overview

Copilot reviewed 5 out of 6 changed files in this pull request and generated no new comments.

[PaddlePaddle-bot]

🤖 AI Code Review | 2026-04-21 19:09:21

📋 Review 摘要

PR 概述：将所有 requirements 文件中已废弃的 pynvml 替换为 nvidia-ml-py，并同步更新文档示例输出。
变更范围：CI 依赖配置、文档
影响面 Tag：CI Docs

---

问题

级别	文件	概述
🟡 建议	fastdeploy/collect_env.py:95,110	collect_env.py 中 pip/conda 包名搜索模式未同步更新

---

🟡 建议：fastdeploy/collect_env.py 中的包名搜索模式需同步更新

fastdeploy/collect_env.py 的 DEFAULT_CONDA_PATTERNS（第 95 行）和 DEFAULT_PIP_PATTERNS（第 110 行）中仍然使用 "pynvml" 作为搜索模式：

# fastdeploy/collect_env.py (第 92–111 行，未修改)
DEFAULT_CONDA_PATTERNS = {
    ...
    "pynvml",   # ← 第 95 行
}

DEFAULT_PIP_PATTERNS = {
    ...
    "pynvml",   # ← 第 110 行
}

影响：切换为 nvidia-ml-py 后，pip list / conda list 中显示的包名是 nvidia-ml-py 而非 pynvml（尽管 import pynvml 仍然有效）。collect-env 工具将无法检测到该包，导致环境诊断输出中缺少 nvidia-ml-py 的版本信息。文档 diff 中已经把 pynvml==12.0.0 的示例行删除，说明这个包名变化已被预料到，但 collect_env.py 未跟进。

建议修复：

DEFAULT_CONDA_PATTERNS = {
    ...
    "nvidia-ml-py",   # 替换 pynvml
}

DEFAULT_PIP_PATTERNS = {
    ...
    "nvidia-ml-py",   # 替换 pynvml
}

---

总体评价

依赖替换本身方向正确且改动简洁，import pynvml 调用无需修改的说明也清晰。唯一需要跟进的是 fastdeploy/collect_env.py 中遗漏的包名搜索模式更新，否则 collect-env 命令在新环境中将无法正常展示 nvidia-ml-py 的版本信息。

[ooooo-create]

@PaddlePaddle-bot nvidia 已经包含了 nvidia-ml-py

[PaddlePaddle-bot]

🤖 Paddle-CI-Agent | ci_status_monitor | 2026-04-27 11:06:30

CI报告基于以下代码生成（30分钟更新一次）:

• PR commit: 1055eaa
• CI 详情

---

1 任务总览

所有 Required 任务均已通过 ✅，PR 可合并。存在 1 个可选任务失败，不阻塞合并，仅供参考。

总执行（含 rerun）	总任务	✅ 通过	❌ 失败	⏳ 运行中	⏸️ 等待中	跳过
0	37	36	1	0	0	0

2 任务状态汇总

2.1 Required 任务：10/10 通过

必选任务阻塞合并，失败需优先处理。

状态	任务	耗时	重跑	根因	修复建议	日志
✅	其余 10 个必选任务通过	-	-	-	-	-

2.2 可选任务 — 26/27 通过

可选任务不阻塞合并，失败仅供参考。

状态	任务	耗时	重跑	日志
❌	CI_HPU	1h4m	-	查看
✅	其余 26 个可选任务通过	-	-	-

---

3 失败详情（仅 required）

无 required 失败任务。