Fine-tuning 6-Billion GPT-J (& other models) with LoRA and 8-bit compression
Primary LanguageJupyter Notebook