#
DIFF.BLOG
New
Following
Discover
Jobs
More
Top Writers
Suggest a blog
Upvotes plugin
Report bug
Contact
About
Sign up
The largest independent dev blog feed.
We surface the best developer writing from thousands of independent blogs, updated daily. The open web is worth fighting for.
Join now
→
Learn more
TOPICS
Gaussian distributed weights for LLMs
1
·
John Cook
·
April 18, 2026, 3:01 p.m.
AI
Number systems
Machine Learning
Quantization
Data Types
LLMs
Summary
This blog post discusses different 4-bit floating point formats for machine learning model weights, specifically NF4 and FP4, and their relevance to downloading quantized weights from Hugging Face.
Read full post on www.johndcook.com →
MORE POSTS LIKE THIS
Building intuition about LLM parameter counts
gpjt ·
Jul 10, 2026
GPT-2
jax
LLM: Learningful Links and Musings
cmart's blog ·
Jul 6, 2026
AI
LLMs
What Is Software, and Will LLMs Replace It?
Federico Tomassetti ·
Jun 23, 2026
AI
Language Engineering
llama.cpp vs. vLLM: Choosing the right local LLM inference engine
Red Hat ·
Jun 15, 2026
AI Inference Engines
llama.cpp
Don't let the LLM speak, just probe it.
Blog J11y ·
Jun 10, 2026
Machine Learning
Natural Language Processing
Quantizing Ideogram 4.0 onto a 3090: an INT8 build that matches FP8 and a 4-bit GGUF that beats NF4
transformerlab ·
Jun 9, 2026
ml-research
Quantization
Discover more posts →
AUTHOR
RECENT POSTS FROM THE AUTHOR
Choose how you want to continue.
Continue with GitHub
Continue with Google