Quantization from the ground up | ngrok blog
ngrok.com · 6,054 words · saved by 3 readers
A complete guide to what quantization is, how it works, and how it's used to compress large language models
Qwen-3-Coder-Next is an 80 billion parameter model 159.4GB in size. That’s roughly how much RAM you would need to run it, and that’s before thinking about long context windows. This is not considered a big model. Rumors have it that frontier models have over 1 trillion parameters, which would require at least 2TB of RAM. The last time I saw that much RAM in one machine was never. But what if I told you we can make LLMs 4x smaller and 2x faster, enough to run very capable models on your laptop, all while losing only 5-10% accuracy. That’s the magic of quantization. What makes large…
saved by
related reading
- A Guide to Quantization in LLMs | Symbl.aisymbl.ai
- Efficient LLM inferencefinbarrtimbers.substack.com
- Reiner Pope – The math behind how LLMs are trained and serveddwarkesh.com
- On neural scaling and the quanta hypothesisericjmichaud.com
- A Visual Guide to Quantization - by Maarten Grootendorstnewsletter.maartengrootendorst.com
- Getting Caught Up to Modern LLM Research | Samarth Goeldev.samarthgoel.com
- 2305.14314arxiv.org
- Do Qwen3.6 27B quantizations break the pelican?quesma.com
- Compression and Intelligencegreene.sh
- TurboQuant: Redefining AI efficiency with extreme compressionresearch.google
- LLM Inference Performance Engineering: Best Practices | Databricks Blogdatabricks.com
- Quantization and Hardware Architecture Co-Design for Matrix-Vector Multiplications of Large Language Modelsieeexplore.ieee.org