I trained a 102M recursive BitNet-v2 model from scratch: 64K context, trained on less than 5B tokens

🌐 链接: https://www.reddit.com/r/LocalLLaMA/comments/1x34nyx/i_trained_a_102m_recursive_bitnetv2_model_from/
🏷️ 分组: Reddit｜AI·自托管·VPS
🕒 时间: 2026-10-11 18:19:08
Reddit
From the LocalLLaMA community on Reddit: I trained a 102M recursive BitNet-v2 model from scratch: 64K context, trained on less…
From the LocalLLaMA community on Reddit: I trained a 102M recursive BitNet-v2 model from scratch: 64K context, trained on less…
Explore this post and more from the LocalLLaMA community