Skip to content

Revert "del split func for fused_experts"#4506

Open
xxyux wants to merge 1 commit into
developfrom
revert-4501-batched_ns
Open

Revert "del split func for fused_experts"#4506
xxyux wants to merge 1 commit into
developfrom
revert-4501-batched_ns

Conversation

@xxyux
Copy link
Copy Markdown
Collaborator

@xxyux xxyux commented May 21, 2026

Reverts #4501

@paddle-bot
Copy link
Copy Markdown

paddle-bot Bot commented May 21, 2026

Thanks for your contribution!

@Paddle-CI-Bot
Copy link
Copy Markdown

PaddleFormers Log Analysis

Run #26246563277 · Attempt 1

日志分析报告

流水线名称 问题标签 修复建议
CI_ILUVATAR 环境问题(container failed) 天数机器自定义容器实现执行失败,属于 Runner 基础设施异常,与 PR 代码无关,建议 Rerun 或联系 CI 维护人员检查 iluvatar-gpu-2 Runner 的容器配置。

失败的测试 case:

无具体测试 case 执行失败。
流水线在 "Checkout code" 阶段即中断,未进入任何测试执行步骤。
错误信息:
  ##[error]Error: failed to run script step: [object Object]
  ##[error]Process completed with exit code 1.
  ##[error]Executing the custom container implementation failed. Please contact your self hosted runner administrator.

根本原因分析:

PR #4506revert-4501-batched_ns)是对 #4501 的 Revert,仅修改了 glm_moe_dsa/modeling.pyminimax_m2/modeling.py_ffn_gate_up 函数的张量 split/concat 逻辑,与本次 CI 失败完全无关

失败发生在 Checkout codeactions/checkout@v4)阶段,Runner iluvatar-gpu-2-nczzk-runner-frh5w 上的 自定义 Kubernetes 容器实现(/home/runner/k8s/index.js)启动失败,导致 checkout step 无法在容器内执行。错误信息 failed to run script step: [object Object] 是 Node.js 运行时层面的容器调度错误,属于基础设施层问题。

修复建议:

  1. 直接 Rerun:点击 GitHub Actions 页面 Re-run job,大概率可恢复(容器调度偶发性失败)。
  2. 若 Rerun 仍失败:联系 CI 维护人员检查 iluvatar-gpu-2 节点上 k8s/index.js 自定义容器实现的状态,确认 Docker image ccr-2vdh3abv-pub.cnc.bj.baidubce.com/device/paddle-ixuca:3.3.0 是否可正常拉取及启动。
  3. PR 代码本身无需修改,等待 CI 环境恢复后重新触发即可。

🔄 每次 Re-run 后自动更新

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants