Show HN: AI/ML benchmark for local LLM inference and XGBoost training on GPU/CPU

Category: devtools

Tags: benchmarking, gpu, llm, xgboost, python

Score: 5.8/10 (Innovation: 4, Technical: 6, Documentation: 7, Utility: 6)

This project provides a unified benchmark suite for measuring local GPU/CPU performance on AI/ML workloads, specifically LLM inference via Ollama and XGBoost training on the HIGGS dataset. It is interesting because it offers reproducible, one-command benchmarks with auto-generated reports and a growing community dashboard, enabling users to compare their hardware against reference systems.

Target audience: data engineers, ML engineers, hardware enthusiasts

Repository: https://github.com/albedan/ai-ml-gpu-bench · Python · MIT · 13 stars

View on Hacker News