MakazhanAlpamys/Soup

Why this repo matters

• •Enables fine-tuning LLMs on consumer GPUs with low VRAM.
• •Simplifies LLM training via a single YAML configuration.
• •Supports multiple fine-tuning methods like DPO, SFT, and LoRA.