Comfy-Org/ComfyUI · error · ValueError

{}\n\nFile path: {}\n\nThe safetensors file is corrupt/incom

Error message

{}\n\nFile path: {}\n\nThe safetensors file is corrupt/incomplete. Check the file size and make sure you have copied/downloaded it correctly.

What it means

When loading a .safetensors file, safetensors raises MetadataIncompleteBuffer when the declared header size exceeds the bytes actually present in the file — i.e. the file is truncated relative to its own header. ComfyUI catches this in load_torch_file and raises a ValueError pointing at incomplete copy/download.

Source

Thrown at comfy/utils.py:148

                if not return_metadata:
                    metadata = None
            else:
                with safetensors.safe_open(ckpt, framework="pt", device=device.type) as f:
                    sd = {}
                    for k in f.keys():
                        tensor = f.get_tensor(k)
                        if DISABLE_MMAP:  # TODO: Not sure if this is the best way to bypass the mmap issues
                            tensor = tensor.to(device=device, copy=True)
                        sd[k] = tensor
                    if return_metadata:
                        metadata = f.metadata()
        except Exception as e:
            if len(e.args) > 0:
                message = e.args[0]
                if "HeaderTooLarge" in message:
                    raise ValueError("{}\n\nFile path: {}\n\nThe safetensors file is corrupt or invalid. Make sure this is actually a safetensors file and not a ckpt or pt or other filetype.".format(message, ckpt))
                if "MetadataIncompleteBuffer" in message:
                    raise ValueError("{}\n\nFile path: {}\n\nThe safetensors file is corrupt/incomplete. Check the file size and make sure you have copied/downloaded it correctly.".format(message, ckpt))
            raise e
    else:
        torch_args = {}
        if MMAP_TORCH_FILES:
            torch_args["mmap"] = True

        pl_sd = torch.load(ckpt, map_location=device, weights_only=True, **torch_args)

        if "state_dict" in pl_sd:
            sd = pl_sd["state_dict"]
        else:
            if len(pl_sd) == 1:
                key = list(pl_sd.keys())[0]
                sd = pl_sd[key]
                if not isinstance(sd, dict):
                    sd = pl_sd
            else:
                sd = pl_sd

View on GitHub (pinned to 1c6d8d45b3)

Solutions

1. Re-download or re-copy the safetensors file and verify its byte size matches the source
2. Check hash if the source publishes one (sha256/civitai hash)
3. If it is a cloud-sync placeholder, force the sync app to materialize the file before loading

Example fix

// before
sd = utils.load_torch_file('model.safetensors')  # truncated

# after
# re-download, then verify expected size before loading
assert os.path.getsize('model.safetensors') == EXPECTED_BYTES
sd = utils.load_torch_file('model.safetensors')

import os
expected = 6425467122  # from source metadata if known
if path.endswith('.safetensors') and os.path.getsize(path) < 1_000_000:
    raise ValueError('safetensors suspiciously small; download likely incomplete')

except ValueError as e:
    if 'corrupt/incomplete' in str(e):
        # trigger re-download path
        raise

When it happens

Trigger: load_torch_file on a .safetensors file truncated by an interrupted download/transfer, or a file corrupted by disk issues, such that the header region is incomplete.

Common situations: Browser download cut off; scp/rsync interrupted; cloud-sync placeholder files; file copied while still being written.

Related errors

• {}\n\nFile path: {}\n\nThe safetensors file is corrupt or in
• Unsupported RIFE model: expected 5 blocks, found {len(channe
• The provided model does not have a heatmap_head. Please use
• INVALID_TAG_FILTER
• ERROR: audio encoder file is invalid or unsupported embed_di

AI-assisted analysis of Comfy-Org/ComfyUI@1c6d8d45b3 (2026-08-14). Data as JSON: /api/errors/a35a4b0bf2841b41. Report an issue: GitHub.