Skip to content

🧠 两小时从 0 训出你的 64M 大模型。Train your own 64M LLM from scratch in 2 hours.

Commits on master

4b86522
[update] trim redundant comments
jingyaogong committed 2 months ago
1e6e909
b00be3e
[fix] unify Ctrl+C handling
jingyaogong committed 2 months ago
cc312c1
e2ee878
[fix] restore top-1 router gradient #858
jingyaogong committed 2 months ago
b63a956
1c87643
fix: eval_llm.py Ctrl+C 退出异常修复
wbbeyourself committed 2 months ago
9fa620e
[update] trim verbose ddp comments
jingyaogong committed 2 months ago
7a9137d
37ac206
[update] simplify safe math eval
jingyaogong committed 2 months ago
21ec325
cdf0a93
9c088dd
[fix] keep agent rollout token-aligned #850
jingyaogong committed 2 months ago
a3c7b01