JosephLi0419 / Rookie-LLM-trainer

Rookie-LLM-trainer

• The repository is aimed at new learners interested in Large Language Models who want to train their own models.

Prompt

• Fewshot learning
• Chain-of-Thought (COT)

Fine tune

• LLama Factory

Evaulation

• lm-evaluation-harness
• GPT4 judgement

Quantization

• gguf: llama.cpp
• AWQ

Other skills

• Merge: mergekit
• Count tokens

Interface:

• Gradio