fix(pt): remove meaningless error raising

[OutisLi]

Summary by CodeRabbit

• Bug Fixes
  • Enhanced distributed training configuration handling to automatically adjust advanced optimization settings when they're enabled without proper distributed training setup, allowing training to proceed instead of failing.

[coderabbitai]

📝 Walkthrough

Walkthrough

The Trainer constructor now automatically downgrades zero_stage to 0 instead of raising a ValueError when distributed training stage configuration is set without an initialized distributed context. The subsequent validation check for change_bias_after_training is bypassed as a result.

Changes

Cohort / File(s)	Summary
Distributed Training Configuration deepmd/pt/train/training.py	Modified error handling in Trainer constructor to downgrade zero_stage to 0 instead of raising an exception when distributed context is unavailable.

Estimated code review effort

🎯 1 (Trivial) | ⏱️ ~2 minutes

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title accurately describes the main change: removing a ValueError that was raised when zero_stage > 0 without distributed context, which the PR characterizes as 'meaningless' since the code now handles this case.
Docstring Coverage	✅ Passed	Docstring coverage is 100.00% which is sufficient. The required threshold is 80.00%.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[coderabbitai]

Actionable comments posted: 1

🤖 Prompt for all review comments with AI agents

Verify each finding against the current code and only fix it if needed.

Inline comments:
In `@deepmd/pt/train/training.py`:
- Around line 185-186: The code silently resets self.zero_stage to 0 when
self.is_distributed is False; change this to emit a clear warning before
mutating the value so users know their requested ZeRO/FSDP stage was ignored
(e.g., call warnings.warn(...) or self.logger.warning(...) with a message like
"zero_stage X requested but distributed launch not detected; forcing
zero_stage=0"), then set self.zero_stage = 0 as before; update the block
containing self.zero_stage and self.is_distributed to perform the warning first
and ensure the message includes the original requested value and guidance about
using torchrun.

🪄 Autofix (Beta)

Fix all unresolved CodeRabbit comments on this PR:

• Push a commit to this branch (recommended)
• Create a new PR with the fixes

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: Repository UI

Review profile: CHILL

Plan: Pro

Run ID: 23b375ec-8cc0-4f65-ba75-efbeb613411d

📥 Commits

Reviewing files that changed from the base of the PR and between 1a1dc59 and 5081e5b.

📒 Files selected for processing (1)

• deepmd/pt/train/training.py

[Copilot]

Pull request overview

This PR changes the PyTorch Trainer initialization behavior so that requesting training.zero_stage > 0 without an initialized distributed process group no longer raises an error and instead falls back to zero_stage = 0.

Changes:

• Removed the ValueError when zero_stage > 0 but training is not running under distributed initialization.
• Added an implicit fallback that sets self.zero_stage = 0 in that case.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[codecov]

Codecov Report

❌ Patch coverage is 0% with 1 line in your changes missing coverage. Please review.
✅ Project coverage is 80.46%. Comparing base (1a1dc59) to head (5081e5b).
⚠️ Report is 1 commits behind head on master.

Files with missing lines	Patch %	Lines
deepmd/pt/train/training.py	0.00%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #5411      +/-   ##
==========================================
- Coverage   80.47%   80.46%   -0.01%     
==========================================
  Files         820      820              
  Lines       86005    86005              
  Branches     4139     4139              
==========================================
- Hits        69209    69208       -1     
  Misses      15521    15521              
- Partials     1275     1276       +1     

☔ 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.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[njzjz]

This change is reasonable to me.