@inproceedings{bai-etal-2023-atformer,
    title = "{ATF}ormer: A Learned Performance Model with Transfer Learning Across Devices for Deep Learning Tensor Programs",
    author = "Bai, Yang  and
      Zhao, Wenqian  and
      Yin, Shuo  and
      Wang, Zixiao  and
      Yu, Bei",
    editor = "Bouamor, Houda  and
      Pino, Juan  and
      Bali, Kalika",
    booktitle = "Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing",
    month = dec,
    year = "2023",
    address = "Singapore",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/fix-sig-urls/2023.emnlp-main.250/",
    doi = "10.18653/v1/2023.emnlp-main.250",
    pages = "4102--4116",
    abstract = "The training and inference efficiency of ever-larger deep neural networks highly rely on the performance of tensor operators on specific hardware platforms. Therefore, a compilation-based optimization flow with automatic tensor generation and parameter tuning is necessary for efficient model deployment. While compilation-based methods with performance models can provide dynamic and suitable code optimization, they suffer from a large design space exploration with rough measurement accuracy and poor transferability among different hardware platforms. This paper presents ATFormer, a simple yet efficient design with attention-inspired modules to accurately predict the performance of optimized operators by capturing global and long-range dependencies within a complete scheduling space. Compared with state-of-the-arts, ATFormer can predict the optimal implementation of tensor operators to reduce inference time with minimal effort on modern DNN benchmarks. Furthermore, ATFormer with pre-trained parameters can quickly adapt to different workloads and hardware via transfer learning."
}