Model Intelligence Sheet
hacnho/gguf-many-tensors-dos-poc overview
GGUF tensor cardinality memory amplification PoC This repository hosts a benign security research PoC for a GGUF parser memory amplification denial of service.…
Runs locally from ~13.4 MB disk (4 GB VRAM class GPUs with llama.cpp / guIDE).
Repository Files & Downloads
1 GGUF files detected
Direct downloads for local inference
| File | Type | Quantization | Size | Link |
|---|---|---|---|---|
| poc-gguf-many-tensors-200k.gguf | GGUF | GGUF | 13.4 MB | Download |
Model Details
Model README
---
tags:
- security
- proof-of-concept
- gguf
- denial-of-service
---
GGUF tensor-cardinality memory-amplification PoC
This repository hosts a benign security research PoC for a GGUF parser memory-amplification denial of service.
Files
poc-gguf-many-tensors-200k.gguf
Reproduction
Use gguf 0.19.0:
python3 -m venv .venv-gguf-019
. .venv-gguf-019/bin/activate
pip install gguf==0.19.0
/usr/bin/time -v python - <<'PY'
import gguf
r = gguf.GGUFReader("poc-gguf-many-tensors-200k.gguf")
print("tensors", len(r.tensors), "fields", len(r.fields), "data_offset", r.data_offset)
PY
Observed on the research machine:
- the file is only about
14 MB GGUFReadermaterializes200,000tensor entries- peak RSS grows to roughly
1.0 GB
Under a 900 MB virtual-memory cap, the same open path fails with MemoryError.
Safety note
This artifact is a non-executable malformed model-file PoC for parser-stress testing only.
Run hacnho/gguf-many-tensors-dos-poc with guIDE
Download guIDE — the AI-native code editor with local LLM inference and 69 built-in tools.
Source: Hugging Face · Compare models