评估了openPangu模型在昇腾芯片上的量化效果，给出实用的精度选择指南。
An Empirical Study of openPangu Quantization on Ascend NPUs
- 在昇腾910B1上系统测试多种量化方法，统一校准与评估流程。
- 8比特权重量化对两个模型几乎无损，4比特对7B模型仍可用，1B模型性能明显下降。
- 2比特及以下量化普遍失效，极端低比特压缩仍具挑战性，适合模型部署优化者参考。
openPangu模型是私有化与国产大语言模型部署的理想目标，但其在昇腾NPUs上进行激进后训练量化时的鲁棒性尚未被系统刻画。本文在华为昇腾910B1 NPU上，对openPangu 1B和7B模型开展受控的实证研究。评估了包括RTN、GPTQ、AWQ、SmoothQuant、GPTAQ、BiLLM和SliM-LLM在内的代表性权重仅量化与权重激活量化方法，在统一校准与评估协议下完成。在18项评测任务中发现，8比特权重量化对两个模型均基本无损；4比特量化对7B模型仍具实用性，但对1B模型在推理、数学和代码任务上显著恶化；超低精度仍具挑战：多数2比特及二值化设置退化为近随机行为，W4A4 SmoothQuant在评估中产生非有限困惑度。该结果为openPangu量化配置提供了面向NPU的精度映射图，并凸显极端低比特压缩的持续困难。
原文摘要 · Abstract (English)
openPangu models are attractive targets for private and domestic large-language-model deployment, yet their robustness under aggressive post-training quantization on Ascend NPUs has not been systematically characterized. This paper conducts a controlled empirical study of openPangu 1B and 7B models on Huawei Ascend 910B1 NPUs. We evaluate representative weight-only and weight-activation post-training quantization methods, including RTN, GPTQ, AWQ, SmoothQuant, GPTAQ, BiLLM, and SliM-LLM, under a unified calibration and evaluation protocol. Across 18 evaluation tasks, we find that 8-bit weight-only quantization is effectively lossless for both models, while 4-bit quantization remains practical for the 7B model but is visibly more harmful for the 1B model on reasoning, math, and code tasks. Ultra-low precision remains challenging: most 2-bit and binary settings collapse to near-random behavior, and W4A4 SmoothQuant produces non-finite perplexity in our evaluation. These results provide an NPU-oriented accuracy map for selecting openPangu quantization settings and highlight the persistent difficulty of extreme low-bit compression.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。