AI模型精选

657MB 本地思考模型:MiniCPM5-1B 经 Claude 追踪微调

Someone Fine-Tuned OpenBMB’s MiniCPM5-1B on Claude Fable 5 Traces to Ship a 657MB Local Thinking Model

精选理由

有人把 Claude Fable 5 的思维链塞进了 MiniCPM5-1B,搞出个 657MB 的本地推理模型,128K 上下文还能看它怎么想的,部署门槛极低。

AI 摘要

社区开发者用 Claude Fable 5 的推理链数据微调了 OpenBMB 的 MiniCPM5-1B,得到一个仅 657MB 的最小化本地推理模型。该模型支持 128K 上下文,推理过程可见。微调后的能力继承自原始模型,但授权条款在模型卡中未完全明确。

图片来源 · marktechpost
原文 · marktechpost

Someone Fine-Tuned OpenBMB’s MiniCPM5-1B on Claude Fable 5 Traces to Ship a 657MB Local Thinking Model

A community developer fine-tuned OpenBMB's MiniCPM5-1B on Claude Fable 5 traces into a 1B model that runs fully local — a 657MB smallest build, 128K context, and visible reasoning. We verify every spec against the Hugging Face cards, separate what a fine-tune actually inherits from real capability, and flag the licensing question the model card leaves open. The post Someone Fine-Tuned OpenBMB’s MiniCPM5-1B on Claude Fable 5 Traces to Ship a 657MB Local Thinking Model appeared first on MarkTechPost .