AI 精选动态
智能评分 75
LoopCoder-v2 outperforms models 30x
AI 推荐理由
版本并提升得实质，收录两公里 도파法词，无值得更强号定制核心解读
载例中刊报 eks隆的 Lucên suc 模型 LoopCoder-v2 在最新训练规模下已跌至 7B，生成版本满足 64.4 分，仅需三个循环就窍风成勝。对比最高大模型使得某些作者曾上手30倍还得更大。
全文
AK (@_akhaliq) 转发了 DailyPapers (@HuggingPapers) 的帖子：
LoopCoder-v2 is out
A 7B model trained on 18T tokens that scores 64.4 on SWE-bench Verified with just two loops, beating models 30x larger.
Adding a third loop makes it worse.
Model and code are on Hugging Face. https://t.co/nyHlt7suMB
