hf-mem

Community

Estimate VRAM needs for Hugging Face models.

Authoralvarobartt
Version1.0.0
Installs0

System Documentation

What problem does it solve?

hf-mem provides estimates of the VRAM and parameter footprint required to load Safetensors-based Hugging Face models, helping you plan deployment and capacity.

Core Features & Use Cases

  • Estimate inference memory for Transformers, Diffusers and Sentence Transformers models using Safetensors metadata and HTTP Range requests.
  • Support gated/private models via HF_TOKEN and optional KV cache estimation with the --experimental flag for causal and multi-modal architectures.
  • Return results as a readable table or JSON, including per-component memory and optional KV cache details to aid capacity planning.

Quick Start

Use the hf-mem CLI to query a model, e.g.: uvx hf-mem --model-id <org/model-name> uvx hf-mem --model-id <org/model-name> --experimental

Dependency Matrix

Required Modules

None required

Components

Standard package

💻 Claude Code Installation

Recommended: Let Claude install automatically. Simply copy and paste the text below to Claude Code.

Please help me install this Skill:
Name: hf-mem
Download link: https://github.com/alvarobartt/hf-mem/archive/main.zip#hf-mem

Please download this .zip file, extract it, and install it in the .claude/skills/ directory.
View Source Repository

Agent Skills Search Helper

Install a tiny helper to your Agent, search and equip skill from 223,000+ vetted skills library on demand.