Skip to content

Unusually Prediction Scores (2011) in Model Output: Meaning and Explanation? #19

Description

@chloe-LGAI

When I run the following inference code:

images = [cv2.imread("./example/img1.png")] model.extract_molecule_bboxes_from_figures(images)

The prediction results contain a score of 2011, which seems very unusual. Here’s an example of the output:

[[{'category': '[Mol]', 'bbox': (0.03601800900450225, 0.07630863546207435, 0.28064032016008006, 0.33645171090096415), 'category_id': 1, 'score': 2011}, {'category': '[Mol]', 'bbox': (0.03751875937968985, 0.6538262629364098, 0.28114057028514255, 0.9157036255448923), 'category_id': 1, 'score': 2011}, {'category': '[Mol]', 'bbox': (0.37868934467233617, 0.08671435847962994, 0.5467733866933466, 0.19510730657916736), 'category_id': 1, 'score': 2011}, {'category': '[Mol]', 'bbox': (0.37868934467233617, 0.6642319859539654, 0.5467733866933466, 0.7726249340535027), 'category_id': 1, 'score': 2011},

Is there any specific meaning or explanation for why the prediction score is 2011?

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions