[Cherry-Pick][Optimization] Enable text-only deployment for multimodal models

[K11OntheBoat]

Motivation

在部署多模态模型的时候，当开启--deploy-modality 'text' 开关，获得一个干净的纯文runtime. 不会有多余的多模部分来干扰服务的资源和推理性能. 收益: xx 多模态模型在使用后, 纯文 benchamrk，QPS 提升2.5倍.

Modifications

enable_mm 代表模型具有多模态能力. enable_mm_runtime 代表多模态runtime，enable_mm_runtime=false 代表纯文runtime.

Usage or Command

多模态模型起服务带上--deploy-modality 'text'开关.

Accuracy Tests

Base 模型，打开和关闭--deploy-modality 'text' ，纯文请求的输入token和输出token一致.

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!

[CLAassistant]

Thank you for your submission! We really appreciate it. Like many open source projects, we ask that you sign our Contributor License Agreement before we can accept your contribution.

---

liuruian seems not to be a GitHub user. You need a GitHub account to be able to sign the CLA. If you have already a GitHub account, please add the email address used for this commit to your account.
_{You have signed the CLA already but the status is still pending? Let us recheck it.}

[fastdeploy-bot]

🤖 AI Code Review | 2026-04-08 12:17 CST

📋 Review 摘要

PR 概述：为多模态模型提供纯文本部署模式，通过 --deploy-modality 'text' 开关获得干净的纯文本 runtime，提升 QPS 约 2.5 倍。

变更范围：config、engine、worker、attention backend

影响面 Tag：[Optimization] [Engine] [Models]

问题

级别	文件	概述
🟡 建议	fastdeploy/worker/input_batch.py:831	ProposerInputBatch 中存在重复的 tensor 初始化
❓ 疑问	fastdeploy/config.py:1941	直接修改 model_config 属性的设计

总体评价

PR 核心逻辑正确，引入 enable_mm_runtime 和 enable_rope_3d_runtime 属性有效区分了"模型是否支持多模态"与"是否启用多模态 runtime"，测试文件也正确更新了 mock 配置。建议修复 ProposerInputBatch.init_share_inputs 中的重复初始化代码。

[fastdeploy-bot]

🟡 建议 代码中存在重复的 tensor 初始化。attn_mask_offsets、attn_mask_offsets_full 和 attn_mask_offsets_decoder 在同一方法中被初始化了两次（第 817-821 行和第 831-839 行），虽然参数相同不影响功能，但浪费了内存分配和计算资源。

建议删除第 831-839 行的重复初始化代码。

[fastdeploy-bot]

❓ 疑问 直接修改 model_config 的 rope_3d 和 use_3d_rope 属性。虽然这不是 bug，且代码中有对应的日志说明，但直接修改模型配置对象可能不是最佳实践。

更好的方式是让 enable_rope_3d_runtime 属性直接检查 deploy_modality，而不是修改 model_config。这样能避免修改原始配置对象，使逻辑更清晰。

[codecov-commenter]

Codecov Report

❌ Patch coverage is 60.46512% with 17 lines in your changes missing coverage. Please review.
⚠️ Please upload report for BASE (release/2.5@2d6fa35). Learn more about missing BASE report.

Files with missing lines	Patch %	Lines
fastdeploy/config.py	57.14%	4 Missing and 2 partials ⚠️
fastdeploy/worker/input_batch.py	50.00%	5 Missing and 1 partial ⚠️
fastdeploy/engine/common_engine.py	0.00%	1 Missing and 2 partials ⚠️
fastdeploy/engine/async_llm.py	0.00%	0 Missing and 1 partial ⚠️
...ecutor/layers/attention/flash_mask_attn_backend.py	0.00%	1 Missing ⚠️

Additional details and impacted files

@@              Coverage Diff               @@
##             release/2.5    #7234   +/-   ##
==============================================
  Coverage               ?   69.48%           
==============================================
  Files                  ?      390           
  Lines                  ?    54382           
  Branches               ?     8574           
==============================================
  Hits                   ?    37788           
  Misses                 ?    13866           
  Partials               ?     2728           

Flag	Coverage Δ
GPU	69.48% <60.46%> (?)

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.