TY - RPRT TI - Beyond Reconstruction Loss in Post-Training Quantization: Balanced Fitting for Large Vision-Language Models AU - Minchan Kang AU - Kyeonghye Park AU - Seungyeon Sa AU - Seoyoung Cho AU - Daeshik Kim AU - Yucheol Cho PY - 2026 UR - https://arxiv.org/abs/2609.34765 ID - 2609.34765 ER -