Douglas Orr
@douglasorr.bsky.social
ML Researcher | Christian | Father
Reposted by Douglas Orr
Your boss emails you a point in 128-billion-dimensional space. It's Llama 8B in bfloat16. They want it compressed.
What should you do 🤔... quantise to NF4? 🧵
What should you do 🤔... quantise to NF4? 🧵
June 12, 2025 at 11:19 AM
Your boss emails you a point in 128-billion-dimensional space. It's Llama 8B in bfloat16. They want it compressed.
What should you do 🤔... quantise to NF4? 🧵
What should you do 🤔... quantise to NF4? 🧵
Reposted by Douglas Orr
Our latest work uses theory from the '50s to figure out how to design weight quantisation formats for LLM inference.
It's called Optimal Formats for Weight Quantisation and has just hit arXiv.
1/6
It's called Optimal Formats for Weight Quantisation and has just hit arXiv.
1/6
May 22, 2025 at 12:25 PM
Our latest work uses theory from the '50s to figure out how to design weight quantisation formats for LLM inference.
It's called Optimal Formats for Weight Quantisation and has just hit arXiv.
1/6
It's called Optimal Formats for Weight Quantisation and has just hit arXiv.
1/6
Reposted by Douglas Orr
Each month our team writes up summaries and analysis of our favourite ML papers. For December we cover:
The Byte Latent Transformer, Large Concept Models, Memory Layers & Phi-4 — all grouped under the title "Spend Your FLOPs Wisely". Here's our take (🧵)
graphcore-research.github.io/papers-of-th...
The Byte Latent Transformer, Large Concept Models, Memory Layers & Phi-4 — all grouped under the title "Spend Your FLOPs Wisely". Here's our take (🧵)
graphcore-research.github.io/papers-of-th...
December Papers: Spend Your FLOPs Wisely
Welcome to Papers of the Month — Graphcore Research’s effort to bring you our pick of the most interesting ML papers. In December we noted a collection of papers which took innovative approaches to al...
graphcore-research.github.io
January 9, 2025 at 11:00 AM
Each month our team writes up summaries and analysis of our favourite ML papers. For December we cover:
The Byte Latent Transformer, Large Concept Models, Memory Layers & Phi-4 — all grouped under the title "Spend Your FLOPs Wisely". Here's our take (🧵)
graphcore-research.github.io/papers-of-th...
The Byte Latent Transformer, Large Concept Models, Memory Layers & Phi-4 — all grouped under the title "Spend Your FLOPs Wisely". Here's our take (🧵)
graphcore-research.github.io/papers-of-th...
Just finished C-crits, my game jam game for #githubgameoff - an asm-like (not C) programming game. Perhaps a bit niche!
douglasorr.itch.io/c-crits
douglasorr.itch.io/c-crits
C-crits by DouglasOrr
Take charge of an army of miniature robot critters, programmed in the assembly-like language crasm.
douglasorr.itch.io
December 2, 2024 at 9:44 PM
Just finished C-crits, my game jam game for #githubgameoff - an asm-like (not C) programming game. Perhaps a bit niche!
douglasorr.itch.io/c-crits
douglasorr.itch.io/c-crits