# Hugging Face hf-mem

> VRAM estimate via HTTP Range, no download. Needs uvx. Gated models need HF_TOKEN

**Subcategory:** Hugging Face  
**License:** Apache-2.0  
**Source:** https://github.com/huggingface/skills/blob/main/skills/hf-mem/SKILL.md

---

## When it is useful

Estimate VRAM/memory to load Safetensors or GGUF from the Hub via HTTP Range — no weight download. Needs uvx. Gated/private models need HF_TOKEN. Distinct from huggingface-local-models (actually running GGUF).

## The Catch

VRAM estimate via HTTP Range, no download. Needs uvx. Gated models need HF_TOKEN. Distinct from huggingface-local-models. Does not train or host.