# Hugging Face publishes a handbook on MoE training optimization

> Original title: Nouamane thread on it: https://x.com/Nouamanetazi/status/2118412817116369254?a=9bwdj-6fl9ZnIbXKSD4&s=20

- Source: Thomas Wolf
- Published: 2026-11-05T19:57:16.000Z
- Bad Influence: https://badinfluence.news/items/vilsoa1fcnn1tqxlfy8rjsaur
- Original: https://x.com/Thom_Wolf/status/2118432479778669038

## Summary

Hugging Face has published a MoE training handbook, which it says covers everything from profiling to large cluster training. The team says it spent 1/2M GPUh optimizing MoE training, and the blog runs about 30k words with 50+ interactive figures.

## Why it matters

The handbook covers MoE training from profiling through large cluster runs, giving practitioners a detailed, figure-rich reference built on about 500,000 GPU hours of optimization work.

## Post

Nouamane thread on it: https://x.com/Nouamanetazi/status/2118412817116369254?a=9bwdj-6fl9ZnIbXKSD4&s=20

## Quoted post @Nouamanetazi

> 📖 The Ultimate MoE Training Handbook 📖
> 
> We spent 1/2M GPUh to optimize MoE training and wrote the most comprehensive blog about it. With 30k words and 50+ interactive figures it covers everything from profiling to large cluster training.
> https://huggingface.co/spaces/huggingface/moe-blogpost

https://x.com/Nouamanetazi/status/2118412817116369254