UkisAI Swift Series / 27B, Flash Next and Bonsai 2 + GSQ-RCO / -63.4% thinking, x1.95 speed with xhigh accuracy
英文标题：UkisAI Swift Series / 27B, Flash Next and Bonsai 2 + GSQ-RCO / -63.4% thinking, x1.95 speed with xhigh accuracy
Hey everyone, Jovan from UkisAI here! Today, we are introducing Swift, a family of efficient reasoning LLMs based on Qwen, trained by penalizing tokens related to pathological overthinking patterns and restoring accuracy via RL (GSPO) and OPD. After amazing feedback and 350k+ downloads in 13 d
打开原文相关 5 · 机会 0
中文
Hey everyone, Jovan from UkisAI here! Today, we are introducing Swift, a family of efficient reasoning LLMs based on Qwen, trained by penalizing tokens related to pathological overthinking patterns and restoring accuracy via RL (GSPO) and OPD. After amazing feedback and 350k+ downloads in 13 d
中文全文翻译正在补齐，可先查看下方英文原文。
English / 原文
Hey everyone, Jovan from UkisAI here! Today, we are introducing Swift, a family of efficient reasoning LLMs based on Qwen, trained by penalizing tokens related to pathological overthinking patterns and restoring accuracy via RL (GSPO) and OPD. After amazing feedback and 350k+ downloads in 13 d