| [feat] 增加DS V4 SFT数据集加载和SFT训练样例配置 | 8 天前 |
| [fix] 修复 swap optimizer checkpoint 保存加载 | 22 天前 |
| Fixed license headers | 1 个月前 |
| [fix] preserve MoE w13 values when exporting HF weights | 1 天前 |
| [feat] DT support scripts | 2 个月前 |
| [feat]support distributed muon optimizer and virtual allocator | 30 天前 |
