benisalla/tiny-llm

A small LLM built from scratch in PyTorch. Loads GPT-2, Llama (SmolLM2), Qwen2.5 and Pythia weights with outputs identical to Hugging Face. LoRA fine-tuning, KV cache, Kaggle/Colab notebook and a Streamlit chatbot.

语言：Python · GitHub 星标：9 · Fork：0

话题：artificial-intelligence · chatbot · deep-learning · fine-tuning · from-scratch · generative-ai · gpt-2 · huggingface-transformers

在 GitHub 上查看原仓库