🚀 New Brumby‑14B‑Base Qwen‑3 variant leverages Power Retention to skip costly full pre‑training. It swaps traditional attention for an efficient transformer‑free design, cutting compute while keeping performance. A promising step for open‑source LLMs and fine‑tuning pipelines. Dive into the details! #Brumby14BBase #Qwen3 #PowerRetention #AttentionFree

🔗 https://aidailypost.com/news/brumby-14b-base-qwen3-variant-uses-power-retention-avoids-full