[API Compatibility] select_scatter/sgn/signbit/slice_scatter/tensordot/tril_indices/triu_indices/vander/logaddexp/logspace/moveaxis/nan_to_num/nanmean/nansum/masked_fill Edit By AI Agent

[zhwesky2010]

PR Category

User Experience

PR Types

Improvements

Description

API Compatibility Edit By AI Agent：
torch.select_scatter
torch.sgn
torch.signbit
torch.slice_scatter
torch.tensordot
torch.tril_indices
torch.triu_indices
torch.vander
torch.logaddexp
torch.logspace
torch.moveaxis
torch.nan_to_num
torch.nanmean
torch.nansum
torch.masked_fill

是否引起精度变化

否

[paddle-bot]

你的PR提交成功，感谢你对开源项目的贡献!
请关注后续CI自动化测试结果，详情请参考Paddle-CI手册。
Your PR has been submitted. Thanks for your contribution!
Please wait for the result of CI firstly. See Paddle CI Manual for details.

[codecov-commenter]

Codecov Report

❌ Patch coverage is 86.48649% with 10 lines in your changes missing coverage. Please review.
⚠️ Please upload report for BASE (develop@1e8221d). Learn more about missing BASE report.

Files with missing lines	Patch %	Lines
python/paddle/tensor/creation.py	53.84%	6 Missing ⚠️
python/paddle/tensor/math.py	91.11%	4 Missing ⚠️

❌ Your patch status has failed because the patch coverage (86.48%) is below the target coverage (90.00%). You can increase the patch coverage or adjust the target coverage.

Additional details and impacted files

@@            Coverage Diff             @@
##             develop   #78971   +/-   ##
==========================================
  Coverage           ?   86.48%           
==========================================
  Files              ?        3           
  Lines              ?       74           
  Branches           ?        0           
==========================================
  Hits               ?       64           
  Misses             ?       10           
  Partials           ?        0           

☔ 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.

[PaddlePaddle-bot]

🤖 Paddle-CI-Agent | pr_review | 2026-06-01 20:41:34

📋 Review 摘要

PR 概述：为多个 Paddle API 添加 PyTorch 风格的参数别名兼容性支持（select_scatter、sgn、signbit、slice_scatter、tensordot、tril/triu_indices、vander、logaddexp、logspace、moveaxis、nan_to_num、nanmean、nansum、masked_fill 等）
变更范围：python/paddle/tensor/、python/paddle/utils/decorator_utils.py、test/legacy_test/
影响面 Tag：[User Experience]

问题

级别	文件	概述
🟡 建议	python/paddle/utils/decorator_utils.py:1264	slice_scatter_decorator 仅处理 int 类型转 list，未处理 PyTorch 允许的 None 值（如 start=None）

📝 PR 规范检查

符合规范。PR Category、PR Types、Description、是否引起精度变化四个必填 section 均存在且内容非空。标题格式 [API Compatibility] 与 diff 影响面一致。

总体评价

PR 整体实现规范，通过装饰器和手动 out 参数支持实现了 PyTorch API 兼容。测试覆盖了多种调用风格（位置参数、关键字参数、别名参数、Tensor 方法）。建议补充 slice_scatter_decorator 对 None 值的处理以提升 PyTorch 兼容度。

[PaddlePaddle-bot]

🟡 建议 isinstance(kwargs[param], int) 仅将 int 类型转为 list，但 PyTorch 的 torch.slice_scatter 允许 start=None（表示从头开始）和 end=None（表示到末尾）。当用户传入 start=None 时，decorator 会将 starts=None 直接传给 slice_scatter，而该函数期望 list 类型，将导致运行时错误。

建议在 int-to-list 转换前增加对 None 的处理逻辑：

int_to_list_params = ['axes', 'starts', 'ends', 'strides']
for param in int_to_list_params:
    if param in kwargs:
        if isinstance(kwargs[param], int):
            kwargs[param] = [kwargs[param]]
        elif kwargs[param] is None:
            del kwargs[param]  # 让函数使用默认行为或后续逻辑处理

或者根据 PyTorch 语义，将 None 转换为对应的默认值（如 start=None → starts=[0]，end=None → ends=[sys.maxsize]）。

[PaddlePaddle-bot]

🤖 Paddle-CI-Agent | ci_status_monitor | 2026-06-01 20:48:08

CI报告基于以下代码生成（30分钟更新一次）:
PR commit: 288bd27 | Merge base: 057b98f (branch: develop)

---

1 Required任务 : 1/1 通过

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

任务	错误类型	置信度	日志
无	—	—	—

2 失败详情

无

[zhwesky2010]

本PR迁移到 #79215