🧠 两小时从 0 训出你的 64M 大模型。Train your own 64M LLM from scratch in 2 hours.
[update] readme
jingyaogong90cd275524f8
diff --git a/README.md b/README.mdindex 914a947..93091a6 100644--- a/README.md+++ b/README.md@@ -1308,9 +1308,11 @@ python eval_toolcall.py --weight agent基于`minimind-3 (64M)`在相同随机种子等超参下的主/客观对比,供参考:-**[A]** minimind-3 (64M, SFT)-**[B]** minimind-3 (64M, GRPO)-**[C]** minimind-3 (64M, Agent-CISPO)+[A] minimind-3 (64M, SFT)++[B] minimind-3 (64M, GRPO)++[C] minimind-3 (64M, Agent-CISPO)### 测试1:主观问答对比@@ -1400,10 +1402,13 @@ agent: 17/20 = 85.00%> 注:以下对比仅为体验参考,非严格 benchmark,样本量有限且带有主观性。-**[A]** minimind-3 (0.06B)-**[B]** minimind-3-moe (0.2B-A0.06B)-**[C]** [baby-llama2-chinese (0.2B)](https://github.com/DLLXW/baby-llama2-chinese)-**[D]** [chatlm-mini-chinese (0.2B)](https://github.com/charent/ChatLM-mini-Chinese)+[A] minimind-3 (0.06B)++[B] minimind-3-moe (0.2B-A0.06B)++[C] [baby-llama2-chinese (0.2B)](https://github.com/DLLXW/baby-llama2-chinese)++[D] [chatlm-mini-chinese (0.2B)](https://github.com/charent/ChatLM-mini-Chinese)### 测试3:问答diff --git a/README_en.md b/README_en.mdindex 20b0089..3b872b5 100644--- a/README_en.md+++ b/README_en.md@@ -1305,9 +1305,11 @@ Let us converge back to the "**unified framework**", reorganizing the table showSubjective/objective comparison based on `minimind-3 (64M)` under the same random seed and other hyperparameters, for reference:-**[A]** minimind-3 (64M, SFT)-**[B]** minimind-3 (64M, GRPO)-**[C]** minimind-3 (64M, Agent-CISPO)+[A] minimind-3 (64M, SFT)++[B] minimind-3 (64M, GRPO)++[C] minimind-3 (64M, Agent-CISPO)### Test 1: Subjective Q&A Comparison@@ -1397,10 +1399,13 @@ So if the task objective is ToolUse, lightweight multi-step calling, and verifia> Note: The following comparison is only for experiential reference, not a strict benchmark; sample size is limited and involves subjectivity.-**[A]** minimind-3 (0.06B)-**[B]** minimind-3-moe (0.2B-A0.06B)-**[C]** [baby-llama2-chinese (0.2B)](https://github.com/DLLXW/baby-llama2-chinese)-**[D]** [chatlm-mini-chinese (0.2B)](https://github.com/charent/ChatLM-mini-Chinese)+[A] minimind-3 (0.06B)++[B] minimind-3-moe (0.2B-A0.06B)++[C] [baby-llama2-chinese (0.2B)](https://github.com/DLLXW/baby-llama2-chinese)++[D] [chatlm-mini-chinese (0.2B)](https://github.com/charent/ChatLM-mini-Chinese)### Test 3: Q&A