Convert bboxes to numpy.ndarray in InstanceData for Detection Task

[Archie2035]

Motivation

The motivation for this PR is to convert the bboxes attribute in the gt_instances object from a list format to numpy.ndarray. This change ensures the data is in a consistent and efficient format for further processing in deep learning models.

Modification

The modification in this PR changes the code related to gt_instances.bboxes and gt_instances.labels assignment from list comprehensions to numpy.ndarray. Specifically, it uses numpy.array to convert the list of bounding boxes and labels into numpy.ndarray.

Modified code:

import numpy as np
from mmdet.structures import InstanceData

if task == 'det':
    gt_instances = InstanceData()
    gt_instances.bboxes = np.array([d['bbox'] for d in data_info['gt_instances']])

BC-breaking (Optional)

There are no backward-compatibility-breaking changes introduced by this modification. The modified code only affects the internal representation of bboxes and labels, but the external interface and expected behavior of the code remain unchanged.

Use cases (Optional)

This PR does not introduce a new feature but refines the existing functionality. However, if there are any internal functions or modules relying on the old list format, ensure they are updated to handle numpy.ndarray appropriately.

Checklist

1. Pre-commit or other linting tools are used to fix the potential lint issues.
2. The modification is covered by complete unit tests. If not, please add more unit test to ensure the correctness.
3. If the modification has potential influence on downstream projects, this PR should be tested with downstream projects, like MMDet or MMPreTrain.
4. The documentation has been modified accordingly, like docstring or example tutorials.