Extraction of every question/answers from the whole Molière's corpus.
Data handling is done in llm_lora
tiny lib.
Try to make an LM/LLM talk like molière
- try_lora.ipynb applies LoRA fine-tuning technics on small LM (that fit into an M3Pro's RAM)
- fine-tune-on-gpt.ipynb prepares a dataset for OpenAI's fine-tuning framework