[XPU] Remove v0 cache schedule and use InputBatch

[cmcamdy]

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]

🤖 AI Code Review | 2026-04-24 02:03:18

📋 Review 摘要

PR 概述：移除 XPU 平台的 V0 KVCache 调度器路径，统一使用 V1 KVCache 调度器，并同步将 pre_ids 缓冲区替换为 token_ids_all。
变更范围：xpu_model_runner、xpu_pre_and_post_process、apply_penalty_multi_scores、sampler、input_batch、spec_decode/mtp
影响面 Tag：XPU KVCache Scheduler Speculative Decoding

---

📝 PR 规范检查

PR 标题包含有效 [XPU] Tag，符合规范。但 Motivation 和 Modifications 章节均未填写，描述完全为空，建议补全以便后续追溯。

描述建议（可直接复制）：

## Motivation
移除 XPU 平台对 V0 KVCache Scheduler 的支持，FastDeploy 2.6 及以上版本 XPU 统一使用 V1 KVCache Scheduler，简化代码路径，减少维护负担。

## Modifications
- `xpu_model_runner.py`：删除 `_init_share_inputs` 方法，改用 `InputBatch` 统一初始化；`insert_prefill_inputs` 替换为 `NotImplementedError` 防止 V0 误用；移除 `_prepare_inputs` 中的 `ENABLE_V1_KVCACHE_SCHEDULER` 条件
- `xpu_pre_and_post_process.py`：删除 `step_xpu` 函数（V0 调度入口）；`update_inputs` 分支统一为 `update_inputs_v1`
- `apply_penalty_multi_scores.py` / `sampler.py`：XPU 路径将 `pre_token_ids` 替换为 `token_ids_all`
- `input_batch.py`：XPU 平台加入 CUDA 同一分支，统一通过 `token_ids_all` 派生 `pre_ids`

---

问题

级别	文件	概述
🟡 建议	xpu_model_runner.py:751	xpu_worker.py 未同步清理，存在死代码分支
🟡 建议	—	未添加单元测试或回归测试

---

总体评价

PR 的核心逻辑清晰正确：通过 InputBatch 统一初始化、删除 step_xpu 函数、将 pre_ids 路径收敛到 token_ids_all，整体方向合理。主要遗漏是 xpu_worker.py 未同步更新（preprocess_new_task 中仍有调用 insert_prefill_inputs 的 else 分支，已成死代码），以及 PR 描述为空。建议补充描述并处理 xpu_worker.py 的遗留分支后合入。

[codecov-commenter]

Codecov Report

❌ Patch coverage is 0% with 1 line in your changes missing coverage. Please review.
⚠️ Please upload report for BASE (develop@029c0d6). Learn more about missing BASE report.

Files with missing lines	Patch %	Lines
fastdeploy/worker/input_batch.py	0.00%	0 Missing and 1 partial ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             develop    #7603   +/-   ##
==========================================
  Coverage           ?   71.73%           
==========================================
  Files              ?      419           
  Lines              ?    57783           
  Branches           ?     9060           
==========================================
  Hits               ?    41448           
  Misses             ?    13512           
  Partials           ?     2823           

Flag	Coverage Δ
GPU	71.73% <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.