AI RESEARCH

LoRaQ: Optimized Low Rank Approximation for 4-bit Quantization

arXiv CS.LG

ArXi:2604.18117v1 Announce Type: new Post-