[XPU] support yiyan model w4a8C8/C16+TP4EP4/PD disaggregation+skip la…

[zccjjj]

…yer mix quant

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)

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!

[PaddlePaddle-bot]

🤖 Paddle-CI-Agent | ci_status_monitor | 2026-05-26 00:16:36

CI报告基于以下代码生成（30分钟更新一次）:

• PR commit: 0c310a5
• Merge base: cc413e0 (branch: develop)
• 查看完整 Diff
• CI 详情

---

1 任务总览

Required 任务 10 个，当前 9 个已通过、1 个失败；阻塞合并的是 Run FastDeploy Unit Tests and Coverage / run_tests_with_coverage，失败原因为差异覆盖率未达到 80% 阈值。

总执行（rerun次数）	总任务	✅ 通过	❌ 失败	⏳ 运行中	⏸️ 等待中	跳过
41(0)	41	35	5	0	1	0

2 任务状态汇总

2.1 Required任务 : 9/10 通过

必选任务阻塞合并，失败需优先处理。

状态	任务	耗时	根因	修复建议	日志	重跑
❌	Run FastDeploy Unit Tests and Coverage / run_tests_with_coverage	1h23m	PR问题：差异覆盖率29%，新增分支未覆盖	补测3个变更文件覆盖未达标行	Job	-
✅	其余 9 个必选任务通过	-	-	-	-	-

2.2 可选任务 — 26/31 通过

可选任务不阻塞合并，失败仅供参考。

状态	任务	耗时	日志	重跑
❌	Run iluvatar Tests / run_iluvatar_cases	2m48s	Job	-
❌	xpu_coverage_report / xpu_coverage_combine	45s	Job	-
❌	Check PR Template	20s	Job	-
❌	Trigger Jenkins for PR	1m58s	Job	-
⏸️	CI_HPU	-	-	-
✅	其余 26 个可选任务通过	-	-	-

---

3 失败详情（仅 required）

Run FastDeploy Unit Tests and Coverage / run_tests_with_coverage — 覆盖率未达标（置信度: 高）

Run FastDeploy Unit Tests and Coverage / run_tests_with_coverage

• 状态: ❌ 失败
• 错误类型: 覆盖率未达标
• 置信度: 高
• 根因摘要: 差异覆盖率29%，新增分支未覆盖
• 分析器: ci_analyze_unittest_fastdeploy

失败用例: 无。日志显示单元测试已通过（TEST_EXIT_CODE=0），失败发生在 Verify Code Coverage Threshold (80%)。

根因详情:
CI 在生成差异覆盖率报告后以 COVERAGE_EXIT_CODE=9 退出，diff_coverage.json 显示本 PR 差异覆盖率为 29%，低于 80% 阈值。未覆盖行集中在本 PR 新增/修改的 fastdeploy/model_executor/layers/quantization/__init__.py、fastdeploy/model_executor/utils.py、fastdeploy/model_executor/layers/moe/moe.py 三个文件，因此判断为 PR 新增逻辑缺少单元测试覆盖。

覆盖率明细:

文件	差异覆盖率	未覆盖行
fastdeploy/model_executor/layers/quantization/__init__.py	50.0%	283, 287
fastdeploy/model_executor/utils.py	37.5%	443, 568, 572, 573, 575
fastdeploy/model_executor/layers/moe/moe.py	16.67%	312-316, 319-321, 357-358

关键日志:

All tests passed
Coverage generation failed (exit code 9)
GPU Patch Coverage Details:
"total_num_lines": 24,
"total_num_violations": 17,
"total_percent_covered": 29,
"num_changed_lines": 154
##[error]Process completed with exit code 9.

修复建议:

1. 在 tests/quantization/test_quantization_init.py 增加 XPU 平台分支用例，mock/patch current_platform.is_xpu() 为 true，验证 get_quantization_config("kvcache") 返回 XPUKvCacheQuantConfig，覆盖 __init__.py L283/L287。
2. 为 fastdeploy/model_executor/utils.py 增加或扩展单测，覆盖 XPU 下 v1_loader_support 对 w4a8 的支持分支（L443），以及 rename_offline_ckpt_suffix_to_fd_suffix 在 MoE w4a8/w4afp8 下将 quant_weight、activation_scale 映射到 weight、in_scale 的分支（L568/L572-L575）。
3. 扩展 tests/layers/test_w4a8_moe.py 中 W4A8 MoE 加载断言，覆盖 _load_in_scale_weight 的 reshape/cast/copy 路径以及 SHARD_ID_TO_SHARDED_DIM 为 None 时进入 in_scale loader 的分支（moe.py L312-L321、L357-L358）。

修复建议摘要: 补测3个变更文件覆盖未达标行

关联变更: fastdeploy/model_executor/layers/quantization/__init__.py L279-L287；fastdeploy/model_executor/utils.py L439-L443、L565-L575；fastdeploy/model_executor/layers/moe/moe.py L310-L321、L355-L358

链接: 查看日志

[codecov-commenter]

Codecov Report

❌ Patch coverage is 5.26316% with 72 lines in your changes missing coverage. Please review.
⚠️ Please upload report for BASE (develop@cc413e0). Learn more about missing BASE report.

Files with missing lines	Patch %	Lines
...cutor/layers/backends/xpu/quantization/kv_cache.py	0.00%	44 Missing ⚠️
fastdeploy/model_executor/layers/moe/moe.py	8.33%	10 Missing and 1 partial ⚠️
...odel_executor/layers/backends/xpu/moe/fused_moe.py	0.00%	6 Missing ⚠️
fastdeploy/model_executor/utils.py	25.00%	5 Missing and 1 partial ⚠️
...loy/model_executor/layers/quantization/__init__.py	25.00%	2 Missing and 1 partial ⚠️
...oy/model_executor/layers/backends/xpu/attention.py	0.00%	2 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             develop    #7924   +/-   ##
==========================================
  Coverage           ?   63.98%           
==========================================
  Files              ?      467           
  Lines              ?    65023           
  Branches           ?     9973           
==========================================
  Hits               ?    41605           
  Misses             ?    20592           
  Partials           ?     2826           

Flag	Coverage Δ
GPU	73.13% <16.66%> (?)
XPU	7.07% <0.00%> (?)

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.