[Cherry-Pick][Others] remove flash_mask (#7819)

[BingooYang]

Motivation

删除flash mask依赖，直接使用paddlefleet中的flashmask

Modifications

删除flash mask依赖包，直接引用flash mask地方增加try catch，并添加相关测试

Usage or Command

NA

Accuracy Tests

NA

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.

[codecov-commenter]

Codecov Report

✅ All modified and coverable lines are covered by tests.
⚠️ Please upload report for BASE (release/2.6@1c5ec61). Learn more about missing BASE report.

Additional details and impacted files

@@              Coverage Diff               @@
##             release/2.6    #8019   +/-   ##
==============================================
  Coverage               ?   71.60%           
==============================================
  Files                  ?      386           
  Lines                  ?    55755           
  Branches               ?     8757           
==============================================
  Hits                   ?    39922           
  Misses                 ?    13017           
  Partials               ?     2816           

Flag	Coverage Δ
GPU	71.60% <100.00%> (?)

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

☔ View full report in Codecov by Harness.
📢 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.

[PaddlePaddle-bot]

🤖 Paddle-CI-Agent | ci_status_monitor | 2026-06-09 06:20:54 UTC+08:00

CI报告基于以下代码生成（30分钟更新一次）:
PR commit: 68b6337 | Merge base: 1c5ec61 (branch: release/2.6)

---

1 Required任务 : 10/10 通过

总执行（rerun次数）	总任务	✅ 通过	❌ 失败	⏳ 运行中	⏸️ 等待中	跳过
53(18)	35	33	2	0	0	0

无

2 失败详情

无

[PaddlePaddle-bot]

🤖 Paddle-CI-Agent | pr_review | 2026-06-09 10:47:57

📋 Review 摘要

PR 概述：移除独立 flash_mask 依赖，并在 SM100 FA4 初始化时改为从 PaddleFleet 命名空间导入 flashmask。
变更范围：fastdeploy/model_executor/layers/attention/flash_attn_backend.py、requirements.txt、tests/layers/test_flash_attn_func.py
影响面 Tag：[OP]

问题

级别	文件	概述
🔴 Bug	fastdeploy/model_executor/layers/attention/flash_attn_backend.py:104	旧 PaddleFleet 命名空间存在但不带 flash_mask 时会跳过可用的新命名空间，导致 FA4 被错误降级

历史 Findings 修复情况

Finding	问题	状态
F1	导入失败后仍继续把未定义的 fa4 写入全局状态	✅ 已修复

📝 PR 规范检查

符合规范。

总体评价

本轮修复已解决上一轮的未定义 fa4 写入问题，但导入回退链仍有一个混合安装场景会误降级 FA4，需要调整为逐候选命名空间尝试后再降级。

[BingooYang]

/re-run all-failed

[zoooo0820]

LGTM

[paddle-bot]

Thanks for your contribution!

[paddle-bot]

Thanks for your contribution!