Data gradients analysis for segmentation of dataset in yolo format

[Soorajkumar3]

💡 Your Question

Is this data gradient not available for segmentation analysis of dataset in yolo format?

[Louis-Dupont]

Hi @Soorajkumar3 We don't provide a built-in class to instantiate a segmentation dataset that would be in YOLO format. But if you already have a dataset class that loads your data (YOLO format or any), then you can use it inside DataGradients.

[Soorajkumar3]

can you share a example , examples in the documentation were not helpful

[Louis-Dupont]

Let's start with that example

from data_gradients.managers.segmentation_manager import SegmentationAnalysisManager 

train_data = ...
val_data = ...
class_names = ...

analyzer = SegmentationAnalysisManager(
    report_title="Testing Data-Gradients Segmentation",
    train_data=train_data,
    val_data=val_data,
    class_names=class_names,
)

analyzer.run()

What you need to have to use this script is:

• training dataset/dataloader train_data
• validation dataset/dataloader val_data
• list of class names that appear in your dataset

This assumes that you already have a dataset/dataloader implementation (as explained in my previous comment). If you don't, then you have two options;

• Either convert your data to COCO or VOC format. We provide built-in dataset formats for both types (Yolo format may come later but we don't support it currently)
• Or build your own dataset class, which should iterate over your images/masks