FlattenQuant: Breaking through the Inference Compute-bound for Large Language Models with Per-tensor Quantization.

Yi Zhang, Fei Yang, Shuang Peng, Fangyu Wang, Aimin Pan

Browse the full COLING paper archive.