Though [Model-Preserving Adaptive Rounding](https://arxiv.org/abs/2505.22988) is published in last week ICML 2026, we all know that this work is more than one year ago, there codes are already there for a long while [Cornell-RelaxML/yaqa-quantization](https://github.com/Cornell-RelaxML/yaqa-quantization).
Though Model-Preserving Adaptive Rounding is published in last week ICML 2026, we all know that this work is more than one year ago, there codes are already there for a long while Cornell-RelaxML/yaqa-quantization.