Cannot check actual batch size at wandb when using autobatch

[Seyoung9304]

Search before asking

• I have searched the YOLOv5 issues and discussions and found no similar questions.

Question

Hi! :)

I found out that when we use autobatch with --batch-size -1 option, -1 is logged at wandb table.

Is there a way to log the actual batch size used for learning in wandb?

Additional

No response

[github-actions]

👋 Hello @Seyoung9304, thank you for your interest in YOLOv5 🚀! Please visit our ⭐️ Tutorials to get started, where you can find quickstart guides for simple tasks like Custom Data Training all the way to advanced concepts like Hyperparameter Evolution.

If this is a 🐛 Bug Report, please provide screenshots and minimum viable code to reproduce your issue, otherwise we can not help you.

If this is a custom training ❓ Question, please provide as much information as possible, including dataset images, training logs, screenshots, and a public link to online W&B logging if available.

For business inquiries or professional support requests please visit https://ultralytics.com or email Glenn Jocher at [email protected].

Requirements

Python>=3.6.0 with all requirements.txt installed including PyTorch>=1.7. To get started:

$ git clone https://github.com/ultralytics/yolov5
$ cd yolov5
$ pip install -r requirements.txt

Environments

YOLOv5 may be run in any of the following up-to-date verified environments (with all dependencies including CUDA/CUDNN, Python and PyTorch preinstalled):

• Google Colab and Kaggle notebooks with free GPU:
• Google Cloud Deep Learning VM. See GCP Quickstart Guide
• Amazon Deep Learning AMI. See AWS Quickstart Guide
• Docker Image. See Docker Quickstart Guide

Status

If this badge is green, all YOLOv5 GitHub Actions Continuous Integration (CI) tests are currently passing. CI tests verify correct operation of YOLOv5 training (train.py), validation (val.py), inference (detect.py) and export (export.py) on MacOS, Windows, and Ubuntu every 24 hours and on every commit.

[glenn-jocher]

@AyushExel do you know if there's a way to update the batch size field in wandb after creation? Or could we delay the wandb logger creation until after AutoBatch is computed here? We can't move this up because it depends on imgsz, which in turn requires model to be defined to get the model stride.

yolov5/train.py

Lines 134 to 141 in c1249a4

	# Image size
	gs = max(int(model.stride.max()), 32) # grid size (max stride)
	imgsz = check_img_size(opt.imgsz, gs, floor=gs * 2) # verify imgsz is gs-multiple
	# Batch size
	if RANK == -1 and batch_size == -1: # single-GPU only, estimate best batch size
	batch_size = check_train_batch_size(model, imgsz)

[AyushExel]

@glenn-jocher yes we can update the batch size in in wandb after the run is created as well. It should be a simple one line fix but I'm away from my system this week :( . I can add this feature on monday

[glenn-jocher]

@AyushExel awesome, thanks for the fast response!

@Seyoung9304 we should be able to push a fix for this next week. Thank you for your patience!

[Seyoung9304]

@AyushExel @glenn-jocher Thank you so much for fast response! 👍 Have a nice day 😀

[glenn-jocher]

@Seyoung9304 good news 😃! Your original issue may now be fixed ✅ in PR #6039 by @AyushExel. To receive this update:

• Git – git pull from within your yolov5/ directory or git clone https://github.com/ultralytics/yolov5 again
• PyTorch Hub – Force-reload model = torch.hub.load('ultralytics/yolov5', 'yolov5s', force_reload=True)
• Notebooks – View updated notebooks
• Docker – sudo docker pull ultralytics/yolov5:latest to update your image

Thank you for spotting this issue and informing us of the problem. Please let us know if this update resolves the issue for you, and feel free to inform us of any other issues you discover or feature requests that come to mind. Happy trainings with YOLOv5 🚀!