LQ-LoRA: Low-rank plus Quantized Matrix Decomposition for Efficient Language Model Finetuning.

Han Guo, Philip Greengard, Eric P. Xing, Yoon Kim

Browse the full ICLR paper archive.