[BugFix] fix cuda-python requirement

[yuanlehome]

Motivation

💡 If this PR is a Cherry Pick, the PR title needs to follow the format by adding the [Cherry-Pick] label at the very beginning and appending the original PR ID at the end. For example, [Cherry-Pick][CI] Add check trigger and logic(#5191)

💡 如若此PR是Cherry Pick，PR标题需遵循格式，在最开始加上[Cherry-Pick]标签，以及最后面加上原PR ID，例如[Cherry-Pick][CI] Add check trigger and logic(#5191)

为支持多cuda版本，取消固定 cuda-python 版本，在使用eplb时，让用户手动下载对应版本的cuda-python

Modifications

Usage or Command

Accuracy Tests

Checklist

• Add at least a tag in the PR title.
  • Tag list: [[FDConfig],[APIServer],[Engine], [Scheduler], [PD Disaggregation], [Executor], [Graph Optimization], [Speculative Decoding], [RL], [Models], [Quantization], [Loader], [OP], [KVCache], [DataProcessor], [BugFix], [Docs], [CI], [Optimization], [Feature], [Benchmark], [Others], [XPU], [HPU], [GCU], [DCU], [Iluvatar], [Metax]]
  • You can add new tags based on the PR content, but the semantics must be clear.
• 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

此PR旨在移除对cuda-python固定版本的依赖，以支持多个CUDA版本。主要变更包括从requirements.txt中删除固定版本，将cuda-python导入改为延迟加载，并在配置检查时添加安装提示。

• 移除requirements.txt中的cuda-python固定版本要求
• 将cuda-python导入从模块级别移至函数内部(延迟导入)
• 在EPLBConfig.check()中添加cuda-python的导入检查和用户友好的错误提示

Reviewed changes

Copilot reviewed 2 out of 3 changed files in this pull request and generated 3 comments.

File	Description
requirements.txt	移除了固定的cuda-python==12.8版本要求，允许用户根据CUDA版本自行安装
fastdeploy/eplb/async_expert_loader.py	将from cuda import cudart从模块顶部移至create_mmap函数内部，实现延迟导入
fastdeploy/config.py	在EPLBConfig.check()中添加cuda-python导入检查，当启用eplb但未安装cuda-python时提供清晰的错误信息和安装指导

[Copilot]

捕获所有异常(Exception)过于宽泛，可能会掩盖真正的错误。建议只捕获ImportError:

try:
    import cuda  # noqa
except ImportError:
    raise ImportError(
        "cuda-python not installed. Install the version matching your CUDA toolkit:\n"
        "  CUDA 12.x → pip install cuda-python==12.*\n"
        "  CUDA 11.x → pip install cuda-python==11.*"
    )

这样可以避免捕获到其他非导入相关的异常(如内存错误、键盘中断等)。

Suggested change

	except Exception:
	except ImportError:

[Copilot]

错误信息缺少如何确定用户CUDA版本的指导。建议添加查看CUDA版本的命令:

raise ImportError(
    "cuda-python not installed. Install the version matching your CUDA toolkit:\n"
    "  Check your CUDA version: nvcc --version or nvidia-smi\n"
    "  CUDA 12.x → pip install cuda-python==12.*\n"
    "  CUDA 11.x → pip install cuda-python==11.*"
)

Suggested change

	"cuda-python not installed. Install the version matching your CUDA toolkit:\n"
	" CUDA 12.x → pip install cuda-python==12.*\n"
	"cuda-python not installed. Install the version matching your CUDA toolkit:\n"
	" Check your CUDA version: nvcc --version or nvidia-smi\n"
	" CUDA 12.x → pip install cuda-python==12.*\n"
	" CUDA 11.x → pip install cuda-python==11.*"

[Copilot]

将from cuda import cudart移到函数内部会破坏现有的单元测试。测试文件tests/eplb/test_async_expert_loader.py第55行正在patchfastdeploy.eplb.async_expert_loader.cudart，但现在cudart是函数内的局部导入，无法在模块级别被patch。

有两个解决方案:

方案1 (推荐): 将导入保留在模块顶部，但使用条件导入处理:

# 在文件顶部
try:
    from cuda import cudart
except ImportError:
    cudart = None

# 在create_mmap函数内
if cudart is None:
    raise ImportError(
        "cuda-python not installed. Install the version matching your CUDA toolkit:\n"
        "  CUDA 12.x → pip install cuda-python==12.*\n"
        "  CUDA 11.x → pip install cuda-python==11.*"
    )

方案2: 修改测试文件中的patch路径为patch("cuda.cudart")，直接mock cuda-python包本身，但这会使测试依赖于cuda-python的存在性。

[codecov-commenter]

Codecov Report

❌ Patch coverage is 18.18182% with 9 lines in your changes missing coverage. Please review.
⚠️ Please upload report for BASE (develop@cb56d46). Learn more about missing BASE report.

Files with missing lines	Patch %	Lines
fastdeploy/config.py	0.00%	4 Missing and 1 partial ⚠️
fastdeploy/eplb/async_expert_loader.py	33.33%	3 Missing and 1 partial ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             develop    #5261   +/-   ##
==========================================
  Coverage           ?   60.54%           
==========================================
  Files              ?      317           
  Lines              ?    38812           
  Branches           ?     5849           
==========================================
  Hits               ?    23500           
  Misses             ?    13456           
  Partials           ?     1856           

Flag	Coverage Δ
GPU	60.54% <18.18%> (?)

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.