Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Feat: Add DOSnet training in PT #3486

Merged
merged 31 commits into from Mar 25, 2024
Merged

Conversation

anyangml
Copy link
Collaborator

@anyangml anyangml commented Mar 18, 2024

This is a follow-up PR on #3452

  • Add DOS loss
  • Fix stat calculation
  • Add UT on training
  • Add e2e JIT test
  • fix dp test data shape

@anyangml
Copy link
Collaborator Author

@iProzd take a look at the loss function, I tried to make it consistent with the TF backend. Need to fix pt/utils/stat.py which is related to #3480

deepmd/pt/loss/dos.py Fixed Show fixed Hide fixed
deepmd/pt/loss/dos.py Fixed Show fixed Hide fixed
@anyangml anyangml marked this pull request as draft March 18, 2024 07:12
Copy link

codecov bot commented Mar 18, 2024

Codecov Report

Attention: Patch coverage is 69.56522% with 42 lines in your changes are missing coverage. Please review.

Project coverage is 77.70%. Comparing base (a58dbc6) to head (3c5e6f5).

Files Patch % Lines
deepmd/pt/loss/dos.py 70.47% 31 Missing ⚠️
deepmd/pt/model/task/dos.py 58.33% 10 Missing ⚠️
deepmd/pt/infer/deep_eval.py 0.00% 1 Missing ⚠️
Additional details and impacted files
@@            Coverage Diff             @@
##            devel    #3486      +/-   ##
==========================================
- Coverage   77.70%   77.70%   -0.01%     
==========================================
  Files         432      433       +1     
  Lines       37356    37490     +134     
  Branches     1620     1620              
==========================================
+ Hits        29028    29132     +104     
- Misses       7467     7497      +30     
  Partials      861      861              

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

@njzjz njzjz linked an issue Mar 19, 2024 that may be closed by this pull request
deepmd/pt/model/task/dos.py Fixed Show resolved Hide resolved
@anyangml anyangml requested review from iProzd, wanghan-iapcm and njzjz and removed request for iProzd and wanghan-iapcm March 19, 2024 10:33
@anyangml anyangml marked this pull request as ready for review March 20, 2024 02:04
@anyangml
Copy link
Collaborator Author

Not sure if it's appropriate to copy test data and examples across examples and tests.

deepmd/pt/infer/deep_eval.py Outdated Show resolved Hide resolved
examples/dos/train/input_torch.json Outdated Show resolved Hide resolved
@anyangml anyangml requested a review from njzjz March 20, 2024 07:47
deepmd/pt/loss/dos.py Outdated Show resolved Hide resolved
@anyangml anyangml requested a review from iProzd March 24, 2024 15:19
@anyangml anyangml added this pull request to the merge queue Mar 25, 2024
@github-merge-queue github-merge-queue bot removed this pull request from the merge queue due to failed status checks Mar 25, 2024
@anyangml anyangml added this pull request to the merge queue Mar 25, 2024
Merged via the queue into deepmodeling:devel with commit 48f06fe Mar 25, 2024
48 checks passed
@anyangml anyangml deleted the feat/dos-train branch March 25, 2024 02:46
@njzjz njzjz mentioned this pull request Apr 2, 2024
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

Successfully merging this pull request may close these issues.

[Feature Request] pt: support DOS in the PyTorch backend
4 participants