Andyyyy64 / whichllm

A CLI tool that auto-detects your hardware and recommends the best local LLM that actually runs on your system, ranked by real benchmarks rather than parameter count.

正常维护 MIT Python Tracked
6.5k 358 15 天前
CIPyPI
localllm

星标趋势

数据积累中,暂无足够数据生成趋势图

AI 分析

项目摘要

whichllm is a Python CLI tool that auto-detects your hardware (GPU/CPU/RAM) and recommends the best local LLM from HuggingFace that fits your system, ranked by real recency-aware benchmarks rather than parameter count. It supports hardware simulation, multi-GPU configurations, and provides actionable recommendations with pasteable output.

为什么值得关注

It solves a real pain point for developers working with local LLMs — choosing a model that actually runs well on their specific hardware — with a simple one-command interface, active development (15 releases in 6 months), and practical features like GPU simulation for hardware planning.

优势

  • Auto-detects GPU/CPU/RAM and recommends models that actually fit
  • Recency-aware benchmarking instead of relying on parameter count
  • Hardware simulation for planning GPU upgrades before purchase
  • Multiple install paths (uvx, pip, brew) and simple CLI interface
  • Active development with 15 releases in 6 months and 23 contributors

局限性

  • No Docker support for containerized environments
  • No examples directory or detailed usage tutorials
  • Limited to GGUF model format based on project topics

使用场景

  • Choosing the best local LLM for your current hardware
  • Planning GPU upgrades by simulating different hardware configurations
  • Comparing model options for a specific system before downloading
  • Sharing model recommendations in markdown format for team discussions
目标用户: Developers and AI/ML practitioners who want to run LLMs locally and need help selecting the right model for their hardware
学习曲线:
分析模型:LongCat-2.0 | 分析时间:1 个月前