[Loader] Add values natural order check to layers grouped validation

[bukejiyu]

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

修复weight_map value 无序导致加载OOM的问题

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 | 2026-04-20 11:37 CST

📋 Review 摘要

PR 概述：在权重加载的 is_layers_are_grouped 判断中增加对 values（文件名）自然序的检查，避免 keys 按 layer 分组但文件名乱序时使用简单迭代器导致加载顺序错误。
变更范围：model_executor/load_weight_utils.py
影响面 Tag：Loader

📝 PR 规范检查

PR 描述中 Motivation 和 Modifications 章节均未填写，建议补充变更动机和修改说明，方便后续维护者理解。

描述建议（可直接复制）：

Motivation

当 model.safetensors.index.json 中的 keys 按 layer 分组、但 values（文件名）未按自然序排列时，safetensors_weights_iterator 按 sorted files 遍历会导致权重加载顺序与预期不一致，需回退到 safetensors_weights_iterator_ordered 逐 key 加载。

Modifications

• 新增 values_are_naturally_ordered 函数，检查 weight_map 的 values 是否按自然序排列。
• 将 is_layers_are_grouped 的判断条件从仅检查 keys 分组改为同时检查 keys 分组和 values 自然序。

问题

级别	文件	概述
🟡 建议	load_weight_utils.py:541	变量命名 is_keys_orders 语义不清晰

总体评价

逻辑正确且有实际意义——补全了原有判断的盲区。变更范围小且安全，仅有一处命名建议。

[PaddlePaddle-bot]

🟡 建议 变量命名 is_keys_orders 语义不够清晰

建议改为 is_keys_grouped 或 are_keys_grouped，与上方的函数名 layers_are_grouped 保持一致，增强可读性：

is_keys_grouped = layers_are_grouped(keys)