1 paper
Junrui Xiao, Zhikai Li, Lianwei Yang +2
Post-training quantization (PTQ) reduces excessive hardware cost by quantizing full-precision models into lower bit representations on a tiny calibration set, without retraining. D…