June 5, 2025 at 11:09 AM
I spent 2 months learning about quantization and am extremely proud of the post I've written about it. I think these are some of the nicest visuals I've ever made, and I love how this compression technique invented in 1898 is being used on the bleeding edge in 2026.
ngrok.com/blog/quantiz...
ngrok.com/blog/quantiz...
Quantization from the ground up | ngrok blog
A complete guide to what quantization is, how it works, and how it's used to compress large language models
ngrok.com
March 25, 2026 at 4:32 PM
I spent 2 months learning about quantization and am extremely proud of the post I've written about it. I think these are some of the nicest visuals I've ever made, and I love how this compression technique invented in 1898 is being used on the bleeding edge in 2026.
ngrok.com/blog/quantiz...
ngrok.com/blog/quantiz...
playing with quantization, dithering, bayer matrices and palettes
January 6, 2025 at 11:49 PM
playing with quantization, dithering, bayer matrices and palettes
We all get to play the game "DCT quantization or pug nose?"
September 11, 2025 at 5:40 PM
We all get to play the game "DCT quantization or pug nose?"
good girls get to keep their floating point weights
bad girls go into the quantization chamber
bad girls go into the quantization chamber
June 5, 2026 at 6:24 PM
good girls get to keep their floating point weights
bad girls go into the quantization chamber
bad girls go into the quantization chamber
Gemma 4 quantization-aware training (QAT) models are now available, bringing AI performance directly to edge devices and consumer GPUs. These checkpoints are optimized with quantization-aware training to dramatically reduce memory requirements and unlock high-speed local inference. 🧵
June 5, 2026 at 4:31 PM
Gemma 4 quantization-aware training (QAT) models are now available, bringing AI performance directly to edge devices and consumer GPUs. These checkpoints are optimized with quantization-aware training to dramatically reduce memory requirements and unlock high-speed local inference. 🧵
and it's even more useless on a video since from averaging multiple frames you can reduce noise and quantization
January 23, 2025 at 8:38 PM
and it's even more useless on a video since from averaging multiple frames you can reduce noise and quantization
Defy quantization; become ungovernable.
July 21, 2024 at 4:14 AM
Defy quantization; become ungovernable.
Am I the last one who can’t forgive autotune and quantization
February 2, 2026 at 4:06 PM
Am I the last one who can’t forgive autotune and quantization
Many people seem to think that quantum mechanics, i.e., “first quantization” is a mystery and that “second quantization” (i.e., quantum field theory) is not (it's a common adage: “1st quantization is a mystery and 2d quantization is a functor”). I disagree! Let's recap: 🧵⤵️ •1/16
December 8, 2025 at 2:57 PM
Many people seem to think that quantum mechanics, i.e., “first quantization” is a mystery and that “second quantization” (i.e., quantum field theory) is not (it's a common adage: “1st quantization is a mystery and 2d quantization is a functor”). I disagree! Let's recap: 🧵⤵️ •1/16
John Clarke, UC Berkeley emeritus professor, has been awarded the 2025 #NobelPrize in #Physics. The @nobelprize.bsky.social committee honored Clarke "for the discovery of macroscopic quantum mechanical tunneling and energy quantization in an electric circuit." news.berkeley.edu/2025/10/07/j...
John Clarke, UC Berkeley emeritus professor, awarded 2025 Nobel Prize in Physics - Berkeley News
The Nobel Prize committee honored Clarke "for the discovery of macroscopic quantum mechanical tunneling and energy quantization in an electric circuit."
news.berkeley.edu
October 7, 2025 at 3:32 PM
John Clarke, UC Berkeley emeritus professor, has been awarded the 2025 #NobelPrize in #Physics. The @nobelprize.bsky.social committee honored Clarke "for the discovery of macroscopic quantum mechanical tunneling and energy quantization in an electric circuit." news.berkeley.edu/2025/10/07/j...
Seriously, the only other time in the course of human events that people stare for so long at FFT quantization artifacts is if they are building a codec.
September 11, 2025 at 5:46 PM
Seriously, the only other time in the course of human events that people stare for so long at FFT quantization artifacts is if they are building a codec.
Color quantization tests using a "staggered" DB-16 palette (32 colors total)
August 4, 2023 at 10:29 AM
Color quantization tests using a "staggered" DB-16 palette (32 colors total)
second quantization
OPENAI DISCOVERS NEW WAY TO CUT INFERENCE COSTS IN HALF – THE INFORMATION
June 30, 2026 at 2:48 PM
second quantization
bro are you ok? your quantization noise spectrum is approaching the perceptibility threshold
November 3, 2023 at 8:32 PM
bro are you ok? your quantization noise spectrum is approaching the perceptibility threshold
Virgin of Fractured Quantization
June 10, 2026 at 9:42 PM
Virgin of Fractured Quantization
it's so nice to have a capable local model i can recommend to people in my life who only have normal laptops, super valuable drop. check out the new prism-ml quantization of Qwen3.8-27b. 6GB weights!
blog: prismml.com/news/bonsai-...
webgpu (browser-native!) demo: huggingface.co/spaces/webml...
blog: prismml.com/news/bonsai-...
webgpu (browser-native!) demo: huggingface.co/spaces/webml...
September 17, 2026 at 10:18 PM
it's so nice to have a capable local model i can recommend to people in my life who only have normal laptops, super valuable drop. check out the new prism-ml quantization of Qwen3.8-27b. 6GB weights!
blog: prismml.com/news/bonsai-...
webgpu (browser-native!) demo: huggingface.co/spaces/webml...
blog: prismml.com/news/bonsai-...
webgpu (browser-native!) demo: huggingface.co/spaces/webml...
📦 NVIDIA / Model-Optimizer
⭐ 3,886 (+22)
🗒 Python
A unified library of SOTA model optimization techniques like quantization, distillation, pruning, neural architecture search, speculative decoding, etc. It compresses deep learning models for downstream deployment frameworks like TensorR...
⭐ 3,886 (+22)
🗒 Python
A unified library of SOTA model optimization techniques like quantization, distillation, pruning, neural architecture search, speculative decoding, etc. It compresses deep learning models for downstream deployment frameworks like TensorR...
GitHub - NVIDIA/Model-Optimizer: A unified library of SOTA model optimization techniques like quantization, distillation, pruning, neural architecture search, speculative decoding, etc. It compresses deep learning models for downstream deployment frameworks like TensorRT-LLM, TensorRT, vLLM, etc. to optimize inference speed.
A unified library of SOTA model optimization techniques like quantization, distillation, pruning, neural architecture search, speculative decoding, etc. It compresses deep learning models for downs...
github.com
September 25, 2026 at 12:17 AM
📦 NVIDIA / Model-Optimizer
⭐ 3,886 (+22)
🗒 Python
A unified library of SOTA model optimization techniques like quantization, distillation, pruning, neural architecture search, speculative decoding, etc. It compresses deep learning models for downstream deployment frameworks like TensorR...
⭐ 3,886 (+22)
🗒 Python
A unified library of SOTA model optimization techniques like quantization, distillation, pruning, neural architecture search, speculative decoding, etc. It compresses deep learning models for downstream deployment frameworks like TensorR...