Skip to content

After I trained on my own 3D data and predicted on the test set, I found some samples with null predictions #77

@maolimin

Description

@maolimin
        "metrics": {
            "1": {
                "Dice": 0.0,
                "FN": 353,
                "FP": 0,
                "IoU": 0.0,
                "TN": 22478495,
                "TP": 0,
                "n_pred": 0,
                "n_ref": 353
            }
        },
        "prediction_file": "/data0/maolimin/U-Mamba/data/nnUNet_results/Dataset202_Breast/nnUNetTrainerUMambaBot__nnUNetPlans__3d_fullres/fold_0/validation/br_130.nii.gz",
        "reference_file": "/data0/maolimin/U-Mamba/data/nnUNet_preprocessed/Dataset202_Breast/gt_segmentations/br_130.nii.gz"

There are also cases where the inference is null on the validation set,Looking forward to your reply

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions