Unsloth

Unsloth: 2x Faster LLM Fine-Tuning with Reduced Memory

Fine-tuning large language models on consumer hardware has been a game of memory optimization Tetris. Every byte of GPU memory is precious — …