# 蚂蚁百灵开源 Ling-3.0 六个基础模型检查点

- 来源：Ant Ling (@AntLingAGI)
- 发布时间：2026-08-19 23:22
- AIHOT 分数：53
- AIHOT 链接：https://aihot.virxact.com/items/cmt08zs1015nerodpvt54a2zu
- 原文链接：https://x.com/AntLingAGI/status/2090097017456590879

## AI 摘要

蚂蚁百灵开源 Ling-3.0-tiny 与 Ling-3.0-flash 的 6 个 Base Model 检查点，覆盖预训练、中期训练及 WSM 合并阶段，均未经过后训练。其用 WSM 替代 LR 衰减进行加权检查点合并，支持持续预训练与离线策略探索；社区可在 tiny-base 上验证策略后扩展至 flash-base。

## 正文

🧵 We’ve open-sourced 6 Base Model checkpoints for Ling-3.0-tiny & Ling-3.0-flash, covering pre-trained, mid-trained, and WSM-merged stages. None has undergone post-training, giving researchers flexible starting points for continued pre-training, fine-tuning, and further research.

Two key highlights:
• We use WSM to replace LR decay with weighted checkpoint merging, making the training process better suited for continual pre-training while enabling offline exploration of different LR decay strategies.
• With one shared training recipe, the community can validate strategies on tiny-base, then scale them to flash-base.