invoke-ai/InvokeAI · error · ValueError

in_channels must be divisible by groups

Error message

in_channels must be divisible by groups

What it means

The PidiNet model's Conv2d wrapper validates group-convolution arguments like nn.Conv2d does: in_channels must be an integer multiple of groups. It raises ValueError('in_channels must be divisible by groups') in __init__ when that invariant is violated, before any weight tensors are created.

Source

Thrown at invokeai/backend/image_util/pidi/model.py:353

            else:
                buffer = torch.zeros(shape[0], shape[1], 5 * 5).to(weights.device)
            weights = weights.view(shape[0], shape[1], -1)
            buffer[:, :, [0, 2, 4, 10, 14, 20, 22, 24]] = weights[:, :, 1:]
            buffer[:, :, [6, 7, 8, 11, 13, 16, 17, 18]] = -weights[:, :, 1:]
            buffer[:, :, 12] = 0
            buffer = buffer.view(shape[0], shape[1], 5, 5)
            y = F.conv2d(x, buffer, bias, stride=stride, padding=padding, dilation=dilation, groups=groups)
            return y
        return func
    else:
        print('impossible to be here unless you force that')
        return None

class Conv2d(nn.Module):
    def __init__(self, pdc, in_channels, out_channels, kernel_size, stride=1, padding=0, dilation=1, groups=1, bias=False):
        super(Conv2d, self).__init__()
        if in_channels % groups != 0:
            raise ValueError('in_channels must be divisible by groups')
        if out_channels % groups != 0:
            raise ValueError('out_channels must be divisible by groups')
        self.in_channels = in_channels
        self.out_channels = out_channels
        self.kernel_size = kernel_size
        self.stride = stride
        self.padding = padding
        self.dilation = dilation
        self.groups = groups
        self.weight = nn.Parameter(torch.Tensor(out_channels, in_channels // groups, kernel_size, kernel_size))
        if bias:
            self.bias = nn.Parameter(torch.Tensor(out_channels))
        else:
            self.register_parameter('bias', None)
        self.reset_parameters()
        self.pdc = pdc

    def reset_parameters(self):

View on GitHub (pinned to 0b6a024f2f)

Solutions

1. Set groups=1 (the default) unless grouped convolutions are specifically required.
2. For depthwise convolutions use groups=in_channels (which trivially divides itself).
3. Adjust in_channels (or the preceding layer's out_channels) so it is an integer multiple of groups.
4. Reduce groups to a divisor of in_channels, e.g. groups = gcd(in_channels, desired_groups).

Example fix

// before
conv = Conv2d(pdc, in_channels=3, out_channels=32, kernel_size=3, groups=2)  # 3 % 2 != 0
// after
conv = Conv2d(pdc, in_channels=4, out_channels=32, kernel_size=3, groups=2)  # or groups=1

if groups > 1 and in_channels % groups != 0:
    raise ValueError(f'in_channels={in_channels} not divisible by groups={groups}')

try:
    conv = Conv2d(pdc, in_channels, out_channels, kernel_size, groups=groups)
except ValueError as e:
    if 'divisible by groups' in str(e):
        groups = 1
        conv = Conv2d(pdc, in_channels, out_channels, kernel_size, groups=groups)
    else:
        raise

When it happens

Trigger: Constructing invokeai/backend/image_util/pidi/model.py Conv2d(pdc, in_channels, out_channels, kernel_size, ..., groups=g) where in_channels % g != 0 — e.g. in_channels=3 with groups=2, or any groups > 1 that does not evenly divide the input channel count.

Common situations: Hand-editing the network to add depthwise (groups=in_channels) or grouped convolutions and miscounting channels; changing the input channel count (e.g. grayscale or 4-channel input) without updating grouped layers; porting layers from another model with different channel widths.

Related errors

• out_channels must be divisible by groups
• Unknown model (%s)
• A submodel type (Tokenizer or TextEncoder) must be provided.
• A submodel type must be provided when loading main pipelines
• Unsupported mask shape: {mask.shape}. Expected (1, h, w) or

AI-assisted analysis of invoke-ai/InvokeAI@0b6a024f2f (2026-08-29). Data as JSON: /api/errors/b89a2aab5de45133. Report an issue: GitHub.