open-mmlab/mmdetection · error · ValueError

Only square RoIs are supporeted in Grid R-CNN

Error message

Only square RoIs are supporeted in Grid R-CNN

What it means

Grid R-CNN's GridHead only supports square RoI feature sizes, so roi_feat_size must be an int (interpreted as both H and W). Passing a tuple like (7, 7) raises ValueError in __init__.

Source

Thrown at mmdet/models/roi_heads/mask_heads/grid_head.py:92

        self.roi_feat_size = roi_feat_size
        self.in_channels = in_channels
        self.conv_kernel_size = conv_kernel_size
        self.point_feat_channels = point_feat_channels
        self.conv_out_channels = self.point_feat_channels * self.grid_points
        self.class_agnostic = class_agnostic
        self.conv_cfg = conv_cfg
        self.norm_cfg = norm_cfg
        if isinstance(norm_cfg, dict) and norm_cfg['type'] == 'GN':
            assert self.conv_out_channels % norm_cfg['num_groups'] == 0

        assert self.grid_points >= 4
        self.grid_size = int(np.sqrt(self.grid_points))
        if self.grid_size * self.grid_size != self.grid_points:
            raise ValueError('grid_points must be a square number')

        # the predicted heatmap is half of whole_map_size
        if not isinstance(self.roi_feat_size, int):
            raise ValueError('Only square RoIs are supporeted in Grid R-CNN')
        self.whole_map_size = self.roi_feat_size * 4

        # compute point-wise sub-regions
        self.sub_regions = self.calc_sub_regions()

        self.convs = []
        for i in range(self.num_convs):
            in_channels = (
                self.in_channels if i == 0 else self.conv_out_channels)
            stride = 2 if i == 0 else 1
            padding = (self.conv_kernel_size - 1) // 2
            self.convs.append(
                ConvModule(
                    in_channels,
                    self.conv_out_channels,
                    self.conv_kernel_size,
                    stride=stride,
                    padding=padding,

View on GitHub (pinned to cfd5d3a985)

Solutions

1. Pass roi_feat_size as an int: roi_feat_size=7 (default)
2. Keep the underlying RoIAlign output square (aligned H==W)

Example fix

# before
roi_feat_size=(7, 7)
# after
roi_feat_size=7

assert isinstance(cfg.get('roi_feat_size', 7), int), 'GridHead roi_feat_size must be int'

def is_int_roi_size(v) -> bool: return isinstance(v, int) and not isinstance(v, bool)

When it happens

Trigger: grid_head=dict(type='GridHead', roi_feat_size=(7, 7)) — supplying a rectangular size tuple as done in some other heads.

Common situations: Reusing bbox_head-style roi_feat_size tuples in a Grid R-CNN config; following generic docs that show tuple roi sizes.

Understand the failure class

Background: "Invalid value" and "allowed values are" config errors: what your library rejected and how to fix it — this error's family across 41 libraries.

Related errors

• grid_points must be a square number
• LoadImageFromFile is not found in the test pipeline
• Visualization needs the "visualizer" termdefined in the conf
• config must be a filename or Config object, but got {type(co
• Unrecognized dataset: {dataset}

AI-assisted analysis of open-mmlab/mmdetection@cfd5d3a985 (2026-08-27). Data as JSON: /api/errors/390d5acdc92d2dc6. Report an issue: GitHub.