Copy wandb param dict before training to avoid overwrites.

[n1mmy]

Copy the hyperparameter dict retrieved from wandb configuration before passing it to train(). Training overwrites parameters in the dictionary (eg scaling obj/box/cls gains), which causes the values reported in wandb to not match the input values. This is confusing as it makes it hard to reproduce a run, and also throws off wandb's Bayesian sweep algorithm.

🛠️ PR Summary

_{Made with ❤️ by Ultralytics Actions}

🌟 Summary

Improved stability in the Weights & Biases (wandb) sweeping process within YOLOv5 repo.

📊 Key Changes

• Modified the method of obtaining hyperparameters (hyp_dict) from the wandb sweep agent to include a .copy() operation.

🎯 Purpose & Impact

• This change ensures that there are no unexpected behaviors during parameter sweeps, as train() modifies parameters that could confuse wandb if the dictionary is not copied.
• Users conducting hyperparameter sweeps with wandb will experience more reliable and consistent logging and parameter handling. 🛠️

[n1mmy]

Here are screenshots showing the issue. Note the value for cls does not match in the before screenshot, but does match in the after screenshot.

[github-actions]

👋 Hello @n1mmy, thank you for submitting a YOLOv5 🚀 PR! To allow your work to be integrated as seamlessly as possible, we advise you to:

• ✅ Verify your PR is up-to-date with upstream/master. If your PR is behind upstream/master an automatic GitHub Actions merge may be attempted by writing /rebase in a new comment, or by running the following code, replacing 'feature' with the name of your local branch:

git remote add upstream https://github.com/ultralytics/yolov5.git
git fetch upstream
# git checkout feature  # <--- replace 'feature' with local branch name
git merge upstream/master
git push -u origin -f

• ✅ Verify all Continuous Integration (CI) checks are passing.
• ✅ Reduce changes to the absolute minimum required for your bug fix or feature addition. "It is not daily increase but daily decrease, hack away the unessential. The closer to the source, the less wastage there is." -Bruce Lee

[AyushExel]

Thanks for the PR. Looks good!

[glenn-jocher]

@n1mmy makes sense, good catch!

PR is merged. Thank you for your contributions to YOLOv5 🚀 and Vision AI ⭐

[n1mmy]

Yay! Thank you both for the quick turnaround and for all your work on this project. I really appreciate it!