Can we use HDF5 Format to read Image and labels for training Image using YOLOV5?

[karndeepsingh]

❔Question

Hi,
Can we use HDF5 Format to read Image and labels for training Image using YOLOV5? As I am trying to run this YOLOV5 model on azure cloud and my team is trying to store annotated image in HDF5 formate. Will that work? or we need pure JPG or PNG files for training.

Additional context

[glenn-jocher]

@karndeepsingh can you provide supporting information, links, references etc for your question?

[glenn-jocher]

@karndeepsingh you may also want to talk to this contributor: #4049

[karndeepsingh]

@glenn-jocher Actually, my team is going to use Databricks platform o train YOLOV5 object detection model. Till now we were training it on locally with less image numbers. But now we are going to train with 100K images with annotation on DataBricks. And they said it stores data in HDF5 format. So,training on Databrick will progress when I use HDF5 format or shall I store images in Azure Blob storage and connect with DataBricks and then load images in PNG OR JPG format from blob storage and train.

Any help and advise would be good.
Thanks

[glenn-jocher]

@karndeepsingh I don't know about other dataset formats, but you can see many examples of YOLOv5 auto-download datasets just by requesting a dataset and letting it download i.e.:

python train.py --data VOC.yaml
python train.py --data VisDrone.yaml
python train.py --data Objects365.yaml
etc...

See data directory for more.

[karndeepsingh]

@glenn-jocher Thanks for adding the information. I had doubt that whether we will be able to train YOLOV5 model on Databricks or not.

Please provide us help on this knowledge as well.

[glenn-jocher]

@karndeepsingh for 3rd party YOLOv5 support you should reach out directly to those authors or organizations.

The standard dataset format is described in our Train Custom Data Tutorial below:

YOLOv5 Tutorials

• Train Custom Data  🚀 RECOMMENDED
• Tips for Best Training Results  ☘️ RECOMMENDED
• Weights & Biases Logging  🌟 NEW
• Supervisely Ecosystem  🌟 NEW
• Multi-GPU Training
• PyTorch Hub  ⭐ NEW
• TorchScript, ONNX, CoreML Export 🚀
• Test-Time Augmentation (TTA)
• Model Ensembling
• Model Pruning/Sparsity
• Hyperparameter Evolution
• Transfer Learning with Frozen Layers  ⭐ NEW
• TensorRT Deployment

[github-actions]

👋 Hello, this issue has been automatically marked as stale because it has not had recent activity. Please note it will be closed if no further activity occurs.

Access additional YOLOv5 🚀 resources:

• Wiki – https://github.com/ultralytics/yolov5/wiki
• Tutorials – https://docs.ultralytics.com/yolov5
• Docs – https://docs.ultralytics.com

Access additional Ultralytics ⚡ resources:

• Ultralytics HUB – https://ultralytics.com
• Vision API – https://ultralytics.com/yolov5
• About Us – https://ultralytics.com/about
• Join Our Team – https://ultralytics.com/work
• Contact Us – https://ultralytics.com/contact

Feel free to inform us of any other issues you discover or feature requests that come to mind in the future. Pull Requests (PRs) are also always welcomed!

Thank you for your contributions to YOLOv5 🚀 and Vision AI ⭐!