저장소
1023097618/train_llm
A small chat model trained from scratch: tokenizer → pretraining → SFT → RL → multimodal.
- Stars
- ★ 20
- Forks
- 0
- Issues
- 0
- Updated
- 11월 4일
- Language
- Python
- License
- 없음
#from-scratch#grpo#llm#multimodal#nanochat#pretraining#pytorch#reinforcement-learning