open-mmlab/mmdetection · error · ValueError

basesize_ratio_range[0] should be either 0.15or 0.2 when inp

Error message

basesize_ratio_range[0] should be either 0.15or 0.2 when input_size is 300, got {basesize_ratio_range[0]}.

What it means

SSDAnchorGenerator auto-computes SSD anchor sizes when min_sizes/max_sizes are not provided. For input_size=300 it only knows two presets selected by the first element of basesize_ratio_range: 0.15 (COCO) or 0.2 (VOC). Any other ratio value has no preset so the constructor rejects it.

Source

Thrown at mmdet/models/task_modules/prior_generators/anchor_generator.py:559

            # calculate anchor ratios and sizes
            min_ratio, max_ratio = basesize_ratio_range
            min_ratio = int(min_ratio * 100)
            max_ratio = int(max_ratio * 100)
            step = int(np.floor(max_ratio - min_ratio) / (self.num_levels - 2))
            min_sizes = []
            max_sizes = []
            for ratio in range(int(min_ratio), int(max_ratio) + 1, step):
                min_sizes.append(int(self.input_size * ratio / 100))
                max_sizes.append(int(self.input_size * (ratio + step) / 100))
            if self.input_size == 300:
                if basesize_ratio_range[0] == 0.15:  # SSD300 COCO
                    min_sizes.insert(0, int(self.input_size * 7 / 100))
                    max_sizes.insert(0, int(self.input_size * 15 / 100))
                elif basesize_ratio_range[0] == 0.2:  # SSD300 VOC
                    min_sizes.insert(0, int(self.input_size * 10 / 100))
                    max_sizes.insert(0, int(self.input_size * 20 / 100))
                else:
                    raise ValueError(
                        'basesize_ratio_range[0] should be either 0.15'
                        'or 0.2 when input_size is 300, got '
                        f'{basesize_ratio_range[0]}.')
            elif self.input_size == 512:
                if basesize_ratio_range[0] == 0.1:  # SSD512 COCO
                    min_sizes.insert(0, int(self.input_size * 4 / 100))
                    max_sizes.insert(0, int(self.input_size * 10 / 100))
                elif basesize_ratio_range[0] == 0.15:  # SSD512 VOC
                    min_sizes.insert(0, int(self.input_size * 7 / 100))
                    max_sizes.insert(0, int(self.input_size * 15 / 100))
                else:
                    raise ValueError(
                        'When not setting min_sizes and max_sizes,'
                        'basesize_ratio_range[0] should be either 0.1'
                        'or 0.15 when input_size is 512, got'
                        f' {basesize_ratio_range[0]}.')
            else:
                raise ValueError(

View on GitHub (pinned to cfd5d3a985)

Solutions

1. Set basesize_ratio_range to (0.2, 0.95) for SSD300 VOC or (0.15, 0.75) for SSD300 COCO
2. Or provide explicit min_sizes and max_sizes lists so the auto-computation branch is skipped
3. If you need a custom ratio at 300, compute your own min/max size lists and pass them directly

Example fix

# before
anchor_generator=dict(
    type='SSDAnchorGenerator',
    input_size=300,
    basesize_ratio_range=(0.1, 0.9))
# after
anchor_generator=dict(
    type='SSDAnchorGenerator',
    input_size=300,
    basesize_ratio_range=(0.2, 0.95))  # SSD300 VOC

bsrr = cfg['anchor_generator']['basesize_ratio_range'][0]
assert cfg['anchor_generator'].get('input_size') != 300 or bsrr in (0.15, 0.2) or 'min_sizes' in cfg['anchor_generator']

When it happens

Trigger: Building SSDAnchorGenerator (or an SSD/Retina SSD config) with input_size=300, min_sizes/max_sizes unset, and basesize_ratio_range whose first element is not exactly 0.15 or 0.2 (e.g. 0.1 or 0.25).

Common situations: Copying an SSD512 config (ratio 0.1) and changing input_size to 300, or tweaking basesize_ratio_range for experiments without supplying explicit min_sizes/max_sizes.

Related errors

• When not setting min_sizes and max_sizes,basesize_ratio_rang
• Only support 300 or 512 in SSDAnchorGenerator when not setti
• Invalid scale {scale}, must be positive.
• Scale must be a number or tuple of int, but got {type(scale)
• Invalid crop_type {crop_type}.

AI-assisted analysis of open-mmlab/mmdetection@cfd5d3a985 (2026-08-27). Data as JSON: /api/errors/3e0d2298ee9596d9. Report an issue: GitHub.