Skip to main content
ModelPilot

ComfyUI "MetadataIncompleteBuffer" / "HeaderTooLarge" loading a safetensors — fix

Also seen as: "MetadataIncompleteBuffer", "HeaderTooLarge", "SafetensorError while deserializing header", "safetensors truncated"

Direct answer

How do I fix "MetadataIncompleteBuffer / HeaderTooLarge (truncated safetensors)"?

This means the .safetensors file is truncated — the download did not finish. safetensors reads an 8-byte header-length prefix and then expects exactly that many bytes; when the file is cut short it fails at load with MetadataIncompleteBuffer or HeaderTooLarge. The file is on disk and looks present, but it is incomplete.

Key facts

  • iThe model download was interrupted or timed out — common on large checkpoints over a flaky link or a CDN redirect.
  • iThe download tool reported success on a partial file (exit codes are unreliable through HuggingFace CDN redirects).
  • iThe disk filled mid-download, leaving a half-written file.
  • iA size-only validator passed a file that is roughly the right size but content-truncated.

Next step

  1. 1Delete the file and re-download it — a truncated safetensors cannot be repaired, and it stays cached until you remove it.
  2. 2Verify the real source size: curl -sI -r 0-0 -L '<url>' returns the full length in Content-Range (the -L matters — HF /resolve/ 302-redirects to a CDN). Compare it to the on-disk byte size.
  3. 3Use a downloader that resumes and validates (aria2c with retries), not a plain fetch that can silently stop.
  4. 4If it keeps truncating, check free disk — a full volume cuts the write short every time.

If that didn't fix it

  • Occasionally the file is complete but corrupt rather than truncated — that surfaces as opaque CUDA errors or NaN output downstream, not MetadataIncompleteBuffer. Same fix: delete and re-download.

Not sure this workflow will run before you rent a GPU?

Paste your ComfyUI workflow JSON into the free checker — it flags known dependencies and an estimated hardware requirement. It does not run the workflow, verify GPU capacity or judge output quality. No signup.

Check my workflow freeSee GPU pricing

Other ComfyUI errors

Fix

GPU Memory Exhausted

Fix

Model Not Found

Fix

Missing Python Package

Fix

Disk Space Exhausted

Fix

Gated Model Access

Fix

missing custom nodes (node type not found)

Frequently Asked Questions

What causes "MetadataIncompleteBuffer / HeaderTooLarge (truncated safetensors)" in ComfyUI?

The model download was interrupted or timed out — common on large checkpoints over a flaky link or a CDN redirect.

How do I fix "MetadataIncompleteBuffer / HeaderTooLarge (truncated safetensors)"?

Delete the file and re-download it — a truncated safetensors cannot be repaired, and it stays cached until you remove it. Verify the real source size: curl -sI -r 0-0 -L '<url>' returns the full length in Content-Range (the -L matters — HF /resolve/ 302-redirects to a CDN). Compare it to the on-disk byte size.