search

mlexchange / mlex_dlsia_segmentation_prototype

Other
3 stars 3 forks source link

Added qlty chopping features and DVC Live #9

Closed TibbersHao closed 8 months ago

TibbersHao commented 8 months ago

This PR covers 2 major feature upgrades:

  1. Added codes for qlty chopping for each frame inside tiled_dataset, and a corresponding collate function to accompany the patch number when passing into DataLoader.
  2. Added DVC Live codes inside train_segmentation() function to save loss and metrics under the unique saving directory, along with the model.pt.

Note: As @xiaoyachong mentioned, upon first time testing, git config might be needed for dvc to function properly. This does not auto commit losses to GitHub repo.

git config --global [user.name](http://user.name/) "John Doe"

git config --global user.email "johndoe@email.com"
gitguardian[bot] commented 8 months ago

️✅ There are no secrets present in this pull request anymore.

If these secrets were true positive and are still valid, we highly recommend you to revoke them. Once a secret has been leaked into a git repository, you should consider it compromised, even if it was deleted immediately. Find here more information about risks.

🦉 GitGuardian detects secrets in your source code to help developers and security teams secure the modern development process. You are seeing this because you or someone else with access to this repository has authorized GitGuardian to scan your pull request.

Our GitHub checks need improvements? Share your feedbacks!

xiaoyachong commented 8 months ago

PR looks good. I left some minor comments. Additionally, I was getting an error from dvc due to the init step, as shown below:

ERROR: failed to initiate DVC - /Users/user/mlex_dlsia_segmentation_prototype/.dvc is ignored by your SCM tool. 
Make sure that it's tracked, for example, by adding '!.dvc' to .gitignore.

Could you please check in with @xiaoyachong and update the init steps in the README file?

Maybe you can try to run 'dvc init' in the repository before training the model.

TibbersHao commented 8 months ago

I forgot to leave a comment last night, new commits ready to be reviewed.

What has been covered:

  1. Changes mentioned in comments above
  2. SMSNet Ensemble method has been ingested into the source code