rohitg00/ai-engineering-from-scratch · error · ValueError

expected 4D input (B,C,H,W), got shape {tuple(x.shape)}

Error message

expected 4D input (B,C,H,W), got shape {tuple(x.shape)}

What it means

Error "expected 4D input (B,C,H,W), got shape {tuple(x.shape)}" thrown in rohitg00/ai-engineering-from-scratch.

Source

Thrown at phases/19-capstone-projects/58-vision-encoder-patches/code/main.py:89

    Output shape on a (B, C, H, W) input is (B, N, hidden) where
    N = (H / patch_size) * (W / patch_size).
    """

    def __init__(self, cfg: FrontEndConfig) -> None:
        super().__init__()
        self.cfg = cfg
        self.proj = nn.Conv2d(
            cfg.in_channels,
            cfg.hidden,
            kernel_size=cfg.patch_size,
            stride=cfg.patch_size,
            bias=True,
        )

    def forward(self, x: torch.Tensor) -> torch.Tensor:
        if x.dim() != 4:
            raise ValueError(f"expected 4D input (B,C,H,W), got shape {tuple(x.shape)}")
        if x.shape[1] != self.cfg.in_channels:
            raise ValueError(
                f"channel mismatch: got {x.shape[1]}, expected {self.cfg.in_channels}"
            )
        if x.shape[2] != self.cfg.image_size or x.shape[3] != self.cfg.image_size:
            raise ValueError(
                f"spatial mismatch: got {tuple(x.shape[2:])}, expected "
                f"({self.cfg.image_size}, {self.cfg.image_size})"
            )
        out = self.proj(x)
        b = out.shape[0]
        out = out.flatten(2).transpose(1, 2)
        return out


class VisionFrontEnd(nn.Module):
    """Patch embed + CLS prepend + 2D sinusoidal position.

View on GitHub (pinned to 39ea8a1c6d)

When it happens

Trigger: Thrown at phases/19-capstone-projects/58-vision-encoder-patches/code/main.py:89 when the library encounters an invalid state.

Common situations: See trigger scenarios.

AI-assisted analysis of rohitg00/ai-engineering-from-scratch@39ea8a1c6d (2026-08-26). Data as JSON: /api/errors/b5558c09c167b022. Report an issue: GitHub.