The model collection of paper: Revisiting Adaptive Rounding with Vectorized Reparameterization for LLM Quantization