🧠 两小时从 0 训出你的 64M 大模型。Train your own 64M LLM from scratch in 2 hours.
[update] dependency version
jingyaogonge84156818aec
diff --git a/requirements.txt b/requirements.txtindex 12790f4..fd2464c 100644--- a/requirements.txt+++ b/requirements.txt@@ -5,17 +5,17 @@ Flask_Cors==4.0.0jieba==0.42.1jsonlines==4.0.0marshmallow==3.22.0-# matplotlib==3.10.0ngrok==1.4.0nltk==3.8numpy==1.26.4openai==1.59.6-# peft==0.7.1psutil==5.9.8+protobuf==6.33.5pydantic==2.11.5rich==13.7.1scikit_learn==1.5.1sentence_transformers==2.3.1+sentencepiece==0.2.0simhash==2.1.2tiktoken==0.10.0transformers==4.57.6@@ -29,4 +29,6 @@ einops==0.8.1swanlab==0.7.11modelscope==1.37.0# torch==2.6.0-# torchvision==0.21.0\ No newline at end of file+# torchvision==0.21.0+# peft==0.7.1+# matplotlib==3.10.0\ No newline at end of file