PBGen：边缘智能的基于反卷积的生成器的部分二值化
PBGen: Partial Binarization of Deconvolution-Based Generators for Edge Intelligence
摘要：
本工作探讨了二值化的反卷积为基础的发电机在GAN的内存节省和加速图像的建设。我们的研究表明，与所有层都可以二值化的卷积神经网络（包括卷积神经网络）不同，生成器中只有部分层可以二值化，而不会造成显著的性能损失。在理论分析和实验验证的基础上，建立了一种基于反卷积操作维数的直接度量方法，该方法可用于快速判定生成器中哪些层可进行二值化。我们的研究结果还表明，生成器和二进制数应该同时进行二进制化，以平衡竞争和更好的性能。基于CelebA的实验结果表明，与原始生成器相比，直接将最先进的二值化技术应用于生成器的所有层将导致2.83 $\times$的性能损失（通过切片Wasserstein距离测量），而将它们应用于选定的层仅可以节省高达25.81 $\times$的内存消耗，在性能损失较小的情况下，推理和训练的加速比分别为1.96 $\times$和1.32 $\times$。
英文摘要：
This work explores the binarization of the deconvolution-based generator in a GAN for memory saving and speedup of image construction. Our study suggests that different from convolutional neural networks (including the discriminator) where all layers can be binarized, only some of the layers in the generator can be binarized without significant performance loss. Supported by theoretical analysis and verified by experiments, a direct metric based on the dimension of deconvolution operations is established, which can be used to quickly decide which layers in the generator can be binarized. Our results also indicate that both the generator and the discriminator should be binarized simultaneously for balanced competition and better performance. Experimental results based on CelebA suggest that directly applying state-of-the-art binarization techniques to all the layers of the generator will lead to 2.83$\times$ performance loss measured by sliced Wasserstein distance compared with the original generator, while applying them to selected layers only can yield up to 25.81$\times$ saving in memory consumption, and 1.96$\times$ and 1.32$\times$ speedup in inference and training respectively with little performance loss.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
