fix(dpmodel): fix `rmse_*` in `more_loss` by njzjz · Pull Request #5105 · deepmodeling/deepmd-kit

njzjz · 2025-12-15T15:13:52Z

The previous value was MSE.

Summary by CodeRabbit

Bug Fixes
- Fixed RMSE metric calculations to correctly display square-root values for energy, force, virial, and atom-related loss measurements, ensuring accurate performance metrics reporting.

_{✏️ Tip: You can customize this high-level summary in your review settings.}

The previous value was MSE.

Copilot

Pull request overview

This PR fixes a bug where RMSE (Root Mean Squared Error) metrics in the more_loss dictionary were incorrectly storing MSE (Mean Squared Error) values. The fix adds xp.sqrt() to convert the L2 loss values to proper RMSE values.

Wraps all l2_*_loss variables with xp.sqrt() when assigning to rmse_* keys in more_loss
Ensures consistency with the existing more_loss["rmse"] calculation at line 268
Applies the fix uniformly across all six RMSE metrics: energy, force, virial, atom energy, prefactor force, and generalized force

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

coderabbitai · 2025-12-15T15:16:56Z

📝 Walkthrough

Walkthrough

Modified deepmd/dpmodel/loss/ener.py to display RMSE metrics as the square root of L2 loss values instead of raw L2 values across energy, force, virial, and related loss components.

Changes

Cohort / File(s)	Summary
RMSE Metric Calculation Correction `deepmd/dpmodel/loss/ener.py`	Applied square-root transformation to L2 loss values for RMSE display across multiple components: energy, force, virial, atom energy, atom preference force, and generalized force. All metrics now properly display root-mean-square error values.

Estimated code review effort

🎯 2 (Simple) | ⏱️ ~10 minutes

Verify mathematical correctness: ensure sqrt is applied consistently to all relevant L2 loss metrics
Confirm no unintended side effects on loss calculation logic or backward compatibility
Check that conditional existence checks via display_if_exist still function as expected with the wrapped values

Pre-merge checks and finishing touches

✅ Passed checks (3 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title directly matches the PR's main objective: fixing rmse_* metrics in more_loss by replacing MSE values with their square-root counterparts.
Docstring Coverage	✅ Passed	Docstring coverage is 100.00% which is sufficient. The required threshold is 80.00%.

✨ Finishing touches

📝 Generate docstrings

🧪 Generate unit tests (beta)

Create PR with unit tests
Post copyable unit tests in a comment

📜 Recent review details

Configuration used: CodeRabbit UI

Review profile: CHILL

Plan: Pro

📥 Commits

Reviewing files that changed from the base of the PR and between b9c82d7 and 0d6fe16.

📒 Files selected for processing (1)

deepmd/dpmodel/loss/ener.py (6 hunks)

🧰 Additional context used

🧠 Learnings (3)

📓 Common learnings

Learnt from: 1azyking
Repo: deepmodeling/deepmd-kit PR: 4169
File: deepmd/pt/loss/ener_hess.py:341-348
Timestamp: 2024-10-05T03:11:02.922Z
Learning: In `deepmd/pt/loss/ener_hess.py`, the `label` uses the key `"atom_ener"` intentionally to maintain consistency with the forked version.

Learnt from: 1azyking
Repo: deepmodeling/deepmd-kit PR: 4169
File: deepmd/pt/loss/ener_hess.py:341-348
Timestamp: 2024-10-08T15:32:11.479Z
Learning: In `deepmd/pt/loss/ener_hess.py`, the `label` uses the key `"atom_ener"` intentionally to maintain consistency with the forked version.

Learnt from: 1azyking
Repo: deepmodeling/deepmd-kit PR: 4169
File: deepmd/utils/argcheck.py:1982-2117
Timestamp: 2024-10-05T03:06:02.372Z
Learning: The `loss_ener_hess` and `loss_ener` functions should remain separate to avoid confusion, despite code duplication.

📚 Learning: 2024-10-08T15:32:11.479Z

Learnt from: 1azyking
Repo: deepmodeling/deepmd-kit PR: 4169
File: deepmd/pt/loss/ener_hess.py:341-348
Timestamp: 2024-10-08T15:32:11.479Z
Learning: In `deepmd/pt/loss/ener_hess.py`, the `label` uses the key `"atom_ener"` intentionally to maintain consistency with the forked version.

Applied to files:

deepmd/dpmodel/loss/ener.py

📚 Learning: 2024-10-05T03:06:02.372Z

Learnt from: 1azyking
Repo: deepmodeling/deepmd-kit PR: 4169
File: deepmd/utils/argcheck.py:1982-2117
Timestamp: 2024-10-05T03:06:02.372Z
Learning: The `loss_ener_hess` and `loss_ener` functions should remain separate to avoid confusion, despite code duplication.

Applied to files:

deepmd/dpmodel/loss/ener.py

🧬 Code graph analysis (1)

deepmd/dpmodel/loss/ener.py (3)

deepmd/dpmodel/loss/loss.py (1)

display_if_exist (40-58)

deepmd/pt/loss/loss.py (1)

display_if_exist (44-54)

deepmd/tf/loss/loss.py (1)

display_if_exist (78-92)

⏰ Context from checks skipped due to timeout of 90000ms. You can increase the timeout in your CodeRabbit configuration to a maximum of 15 minutes (900000ms). (29)

GitHub Check: CodeQL analysis (python)
GitHub Check: Test C++ (false, false, false, true)
GitHub Check: Test C++ (false, true, true, false)
GitHub Check: Test C++ (true, false, false, true)
GitHub Check: Test C++ (true, true, true, false)
GitHub Check: Test Python (1, 3.10)
GitHub Check: Test Python (2, 3.10)
GitHub Check: Test Python (5, 3.10)
GitHub Check: Test Python (5, 3.13)
GitHub Check: Test Python (4, 3.13)
GitHub Check: Test Python (6, 3.10)
GitHub Check: Test Python (6, 3.13)
GitHub Check: Test Python (3, 3.10)
GitHub Check: Test Python (1, 3.13)
GitHub Check: Test Python (3, 3.13)
GitHub Check: Test Python (4, 3.10)
GitHub Check: Test Python (2, 3.13)
GitHub Check: Build C++ (clang, clang)
GitHub Check: Build C++ (rocm, rocm)
GitHub Check: Build C++ (cuda120, cuda)
GitHub Check: Build C library (2.18, libdeepmd_c.tar.gz)
GitHub Check: Build C++ (cpu, cpu)
GitHub Check: Build wheels for cp310-manylinux_aarch64
GitHub Check: Build wheels for cp311-macosx_arm64
GitHub Check: Build wheels for cp311-win_amd64
GitHub Check: Build wheels for cp311-macosx_x86_64
GitHub Check: Build wheels for cp311-manylinux_x86_64
GitHub Check: Analyze (python)
GitHub Check: Analyze (c-cpp)

🔇 Additional comments (1)

deepmd/dpmodel/loss/ener.py (1)

182-184: LGTM! RMSE calculations are now mathematically correct.

All six RMSE metrics (rmse_e, rmse_f, rmse_v, rmse_ae, rmse_pf, rmse_gf) now correctly compute the square root of their respective L2 losses, properly converting MSE to RMSE values. Note that the PyTorch backend (deepmd/pt/loss/ener.py) already implements this correctly with .sqrt() calls, and the TensorFlow backend uses a different post-processing architecture with np.sqrt().

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

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

codecov · 2025-12-15T16:06:56Z

Codecov Report

❌ Patch coverage is 75.00000% with 1 line in your changes missing coverage. Please review.
✅ Project coverage is 82.14%. Comparing base (b9c82d7) to head (0d6fe16).

Files with missing lines	Patch %	Lines
deepmd/dpmodel/loss/ener.py	75.00%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##            devel    #5105      +/-   ##
==========================================
- Coverage   82.14%   82.14%   -0.01%     
==========================================
  Files         709      709              
  Lines       72453    72454       +1     
  Branches     3616     3615       -1     
==========================================
  Hits        59515    59515              
- Misses      11775    11776       +1     
  Partials     1163     1163

☔ 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.

The previous value was MSE.  ## Summary by CodeRabbit * **Bug Fixes** * Fixed RMSE metric calculations to correctly display square-root values for energy, force, virial, and atom-related loss measurements, ensuring accurate performance metrics reporting. <sub>✏️ Tip: You can customize this high-level summary in your review settings.</sub>

fix(dpmodel): fix rmse_* in more_loss

0d6fe16

The previous value was MSE.

Copilot AI review requested due to automatic review settings December 15, 2025 15:13

github-actions bot added the Python label Dec 15, 2025

Copilot started reviewing on behalf of njzjz December 15, 2025 15:14 View session

Copilot AI reviewed Dec 15, 2025

View reviewed changes

njzjz requested a review from wanghan-iapcm December 16, 2025 06:12

wanghan-iapcm approved these changes Dec 16, 2025

View reviewed changes

wanghan-iapcm added this pull request to the merge queue Dec 16, 2025

Merged via the queue into deepmodeling:devel with commit 89a3180 Dec 16, 2025
64 checks passed

coderabbitai bot mentioned this pull request Jan 5, 2026

fix(dpmodel): rmse_e loss normalized by n_atoms #5124

Merged

This was referenced Mar 15, 2026

fix(dpmodel): align Loss.call return type with implementation #5314

Closed

fix(dpmodel): fix natoms[0] bug, einsum, and return type in EnergyLoss #5325

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

fix(dpmodel): fix `rmse_*` in `more_loss`#5105

fix(dpmodel): fix `rmse_*` in `more_loss`#5105
wanghan-iapcm merged 1 commit intodeepmodeling:develfrom
njzjz:fix-dpmodel-loss-rmse

njzjz commented Dec 15, 2025 •

edited by coderabbitai bot

Loading

Uh oh!

Copilot AI left a comment

Uh oh!

coderabbitai bot commented Dec 15, 2025

Walkthrough

Changes

Estimated code review effort

Uh oh!

codecov bot commented Dec 15, 2025 •

edited

Loading

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

Conversation

njzjz commented Dec 15, 2025 • edited by coderabbitai bot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary by CodeRabbit

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

coderabbitai bot commented Dec 15, 2025

Walkthrough

Changes

Estimated code review effort

Pre-merge checks and finishing touches

Uh oh!

codecov bot commented Dec 15, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Codecov Report

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

njzjz commented Dec 15, 2025 •

edited by coderabbitai bot

Loading

codecov bot commented Dec 15, 2025 •

edited

Loading