4-bit floating point FP4
(news.ycombinator.com)
1.
2.
Running local models on Macs gets faster with Ollama's MLX support
(arstechnica.com)
3.
Ollama is now powered by MLX on Apple Silicon in preview
(news.ycombinator.com)
4.
5.
SVDQuant+NVFP4: 4× Smaller, 3× Faster FLUX with 16-bit Quality on Blackwell GPUs
(news.ycombinator.com)