open-mmlab/mmdetection · error · ValueError

out_indices must be a subset of range(0, 8). But received {o

Error message

out_indices must be a subset of range(0, 8). But received {out_indices}

What it means

MobileNetV2.__init__ validates out_indices: every element must be an integer in 0..7 (the backbone has 8 stages of inverted residual blocks). Passing indices outside that set raises ValueError because no feature map exists for them.

Source

Thrown at mmdet/models/backbones/mobilenet_v2.py:82

            warnings.warn('DeprecationWarning: pretrained is deprecated, '
                          'please use "init_cfg" instead')
            self.init_cfg = dict(type='Pretrained', checkpoint=pretrained)
        elif pretrained is None:
            if init_cfg is None:
                self.init_cfg = [
                    dict(type='Kaiming', layer='Conv2d'),
                    dict(
                        type='Constant',
                        val=1,
                        layer=['_BatchNorm', 'GroupNorm'])
                ]
        else:
            raise TypeError('pretrained must be a str or None')

        self.widen_factor = widen_factor
        self.out_indices = out_indices
        if not set(out_indices).issubset(set(range(0, 8))):
            raise ValueError('out_indices must be a subset of range'
                             f'(0, 8). But received {out_indices}')

        if frozen_stages not in range(-1, 8):
            raise ValueError('frozen_stages must be in range(-1, 8). '
                             f'But received {frozen_stages}')
        self.out_indices = out_indices
        self.frozen_stages = frozen_stages
        self.conv_cfg = conv_cfg
        self.norm_cfg = norm_cfg
        self.act_cfg = act_cfg
        self.norm_eval = norm_eval
        self.with_cp = with_cp

        self.in_channels = make_divisible(32 * widen_factor, 8)

        self.conv1 = ConvModule(
            in_channels=3,
            out_channels=self.in_channels,

View on GitHub (pinned to cfd5d3a985)

Solutions

1. Restrict out_indices to values within range(0, 8), e.g. (1, 2, 4, 7)
2. Count the MobileNetV2 stages for your widen_factor config and pick valid exit points
3. Match the number of out_indices to what the downstream neck (FPN) expects

Example fix

// before
backbone=dict(type='MobileNetV2', out_indices=(0, 1, 2, 3, 4, 5, 6, 7, 8))
// after
backbone=dict(type='MobileNetV2', out_indices=(1, 2, 4, 7))

assert set(out_indices).issubset(set(range(0, 8))), 'out_indices must be within 0..7'

When it happens

Trigger: MobileNetV2(out_indices=(8,)) or out_indices=[-1, 4] or out_indices=(0, 1, 2, 3, 4, 5, 6, 7, 8); any subset not within set(range(0, 8)).

Common situations: Configuring FPN/neck outputs to take extra feature levels; copying out_indices from another backbone (e.g. ResNet depth-50 with 4 stages) into MobileNetV2; off-by-one when listing stages.

Related errors

• metric must be a list or a str.
• metrics {iou_metrics} is not supported. Only supports mIoU/m
• the item in out_indices must in range(0, {len(self.layer_set
• pretrained must be a str or None
• frozen_stages must be in range(-1, 8). But received {frozen_

AI-assisted analysis of open-mmlab/mmdetection@cfd5d3a985 (2026-08-27). Data as JSON: /api/errors/2981b4eae86eff16. Report an issue: GitHub.