← Papers

QLoRA: Efficient Finetuning of Quantized LLMs

Dettmers et al. · 2023

Read the paper ↗

My notes

Relevant to Code-Eagle AI's reviewer model. NF4 + paged optimizers is the practical bit.