Event ·Multi-source

Ripple-65B 首个 GGUF 量化版本发布，可本地运行

atomic_chat 团队将 Ripple-65B 移植到 llama.cpp，此前该模型无法在本地运行。移植结果与 RicursiveAI 的参考代码对比，token 选择 100% 一致。目前提供 Q8_0（68.8 GB）和 Q4_K_M（38.8 GB）两种磁盘占用规格，24 GB 和 32 GB 的更小量化版本正在制作中。

Current read

Ripple-65B 终于能在自己电脑上跑了，已经移植到 llama.cpp，还有小显存版本在路上。

3 sources43° AI heatLast updated 2026/11/4 22:58:15

Evidence

3 sources

Primary source@atomic_chat_hq

2026/11/4 22:58:15View source

Related source 1arXiv cs.LG

2026/11/3 12:30:30View source

Related source 2Hugging Face

2026/11/5 11:36:58View source

Conflict check

Dedup data only shows that these sources discuss the same event, not that they agree. There is no structured supporting or refuting evidence yet, so no conflict is inferred automatically.

Open the story