dailyai.report

The new MAGNET framework empowers researchers worldwide to autonomously generate, train, and deploy domain‑specific language models on standard CPUs, eliminating the need for expensive GPUs. By combining an automated research pipeline, low‑bit ternary training, and efficient distributed merging, it accelerates model development across industries.