文章研究了基于Transformer模型的中文文本生成方法,重点探讨了Transformer模型的编码器-解码器结构及其工作原理。在详细分析了编码器和解码器的工作机制后,文章利用Hugging Face Transformers开源模型进行了中文文本生成实验。结果表...文章研究了基于Transformer模型的中文文本生成方法,重点探讨了Transformer模型的编码器-解码器结构及其工作原理。在详细分析了编码器和解码器的工作机制后,文章利用Hugging Face Transformers开源模型进行了中文文本生成实验。结果表明,该方法在自制数据集上取得了良好的效果,其准确率、精确率和召回率分别达到92.5%、91.8%和90.6%。该研究不仅拓展了中文自然语言处理的理论基础,还为实际应用提供了高效的技术支持。展开更多
细粒度图像分类是计算机视觉领域的一大分类任务,其难点在于如何通过类别监督信息自主地找到判别性区域.提出一种新的通道-空间融合注意力模块,基于该模块设计了一种新的Swin Transformer算法SwinT⁃NCSA(a Swin Transformer based on a ...细粒度图像分类是计算机视觉领域的一大分类任务,其难点在于如何通过类别监督信息自主地找到判别性区域.提出一种新的通道-空间融合注意力模块,基于该模块设计了一种新的Swin Transformer算法SwinT⁃NCSA(a Swin Transformer based on a novel channel⁃spatial attention module),分别从通道维和空间维同时提取特征,再将其融入到Swin Transformer模型中以提高其小尺度中多头注意力信息的提取能力.SwinT⁃NCSA算法特别关注了对分类有用的区域,同时忽视对分类无用的背景区域,以此在细粒度图像分类任务中达到较高的分类准确率.在FGVC Aircraft飞机数据集、CUB-200-2011鸟类数据集和Stanford Cars车类数据集3个公共数据集上的实验表明,SwinT⁃NCSA算法可以分别取得93.3%、88.4%和94.7%的准确率,优于同类算法.展开更多
Under the Flaschka-Newell Lax pair,the Darboux transformation for the Painlevé-Ⅱequation is constructed by the limiting technique.With the aid of the Darboux transformation,the rational solutions are represented...Under the Flaschka-Newell Lax pair,the Darboux transformation for the Painlevé-Ⅱequation is constructed by the limiting technique.With the aid of the Darboux transformation,the rational solutions are represented by the Gram determinant,and then we give the large y asymptotics of the determinant and the rational solutions.Finally,the solution of the corresponding Riemann-Hilbert problem is obtained from the Darboux matrices.展开更多
As known, the method to obtain a sequence space by using convergence field of an infinite matrix is an old method in the theory of sequence spaces. However, the study of convergence field of an infinite matrix in the ...As known, the method to obtain a sequence space by using convergence field of an infinite matrix is an old method in the theory of sequence spaces. However, the study of convergence field of an infinite matrix in the space of almost convergent sequences is so new （see [15]）. The purpose of this paper is to introduce the new spaces ^ ～f and fo consisting of all sequences whose Ceshro transforms of order one are in the spaces f and ^ ～ f0, respectively. Also, in this paper, we show that ^ ～f and ^ ～f0 are linearly isomorphic to the spaces f and f0, respectively. The β- and γ-duals of the spaces ^ ～f and 2% are computed. Furthermore, the classes （^ ～f： μ） and （μ ： f） of infinite matrices are characterized for any given sequence space μ, and determined the necessary and sufficient conditions on a matrix A to satisfy Bc-core（Ax） K-core（x）, K-core（Ax） Bg-core（x）, Bc-core（Ax） Be-core（x）, Bc-core（Ax） t-core（x） for all x ∈ t∞.展开更多
The class f of almost convergent sequences was introduced by G.G. Lorentz, using the idea of the Banach limits [A contribution to the theory of divergent sequences, Acta Math. 80（1948）, 167-190]. Let fo（B） and f（...The class f of almost convergent sequences was introduced by G.G. Lorentz, using the idea of the Banach limits [A contribution to the theory of divergent sequences, Acta Math. 80（1948）, 167-190]. Let fo（B） and f（B） be the domain of the double sequential band matrix B（r, s） in the sequence spaces f0 and f. In this article, the β- and γ-duals of the space f（B） are determined. Additionally, we give some inclusion theorems concerning with the spaces f0（B） and f（β）. Moreover, the classes （f（B） ： μ） and （μ： f（B）） of infinite matrices are characterized, and the characterizations of some other classes are also given as an application of those main results, where μ is an arbitrary sequence space.展开更多
文摘文章研究了基于Transformer模型的中文文本生成方法,重点探讨了Transformer模型的编码器-解码器结构及其工作原理。在详细分析了编码器和解码器的工作机制后,文章利用Hugging Face Transformers开源模型进行了中文文本生成实验。结果表明,该方法在自制数据集上取得了良好的效果,其准确率、精确率和召回率分别达到92.5%、91.8%和90.6%。该研究不仅拓展了中文自然语言处理的理论基础,还为实际应用提供了高效的技术支持。
文摘细粒度图像分类是计算机视觉领域的一大分类任务,其难点在于如何通过类别监督信息自主地找到判别性区域.提出一种新的通道-空间融合注意力模块,基于该模块设计了一种新的Swin Transformer算法SwinT⁃NCSA(a Swin Transformer based on a novel channel⁃spatial attention module),分别从通道维和空间维同时提取特征,再将其融入到Swin Transformer模型中以提高其小尺度中多头注意力信息的提取能力.SwinT⁃NCSA算法特别关注了对分类有用的区域,同时忽视对分类无用的背景区域,以此在细粒度图像分类任务中达到较高的分类准确率.在FGVC Aircraft飞机数据集、CUB-200-2011鸟类数据集和Stanford Cars车类数据集3个公共数据集上的实验表明,SwinT⁃NCSA算法可以分别取得93.3%、88.4%和94.7%的准确率,优于同类算法.
基金Project supported by the National Natural Science Foundation of China (Grant No.12101246)。
文摘Under the Flaschka-Newell Lax pair,the Darboux transformation for the Painlevé-Ⅱequation is constructed by the limiting technique.With the aid of the Darboux transformation,the rational solutions are represented by the Gram determinant,and then we give the large y asymptotics of the determinant and the rational solutions.Finally,the solution of the corresponding Riemann-Hilbert problem is obtained from the Darboux matrices.
文摘As known, the method to obtain a sequence space by using convergence field of an infinite matrix is an old method in the theory of sequence spaces. However, the study of convergence field of an infinite matrix in the space of almost convergent sequences is so new （see [15]）. The purpose of this paper is to introduce the new spaces ^ ～f and fo consisting of all sequences whose Ceshro transforms of order one are in the spaces f and ^ ～ f0, respectively. Also, in this paper, we show that ^ ～f and ^ ～f0 are linearly isomorphic to the spaces f and f0, respectively. The β- and γ-duals of the spaces ^ ～f and 2% are computed. Furthermore, the classes （^ ～f： μ） and （μ ： f） of infinite matrices are characterized for any given sequence space μ, and determined the necessary and sufficient conditions on a matrix A to satisfy Bc-core（Ax） K-core（x）, K-core（Ax） Bg-core（x）, Bc-core（Ax） Be-core（x）, Bc-core（Ax） t-core（x） for all x ∈ t∞.
文摘The class f of almost convergent sequences was introduced by G.G. Lorentz, using the idea of the Banach limits [A contribution to the theory of divergent sequences, Acta Math. 80（1948）, 167-190]. Let fo（B） and f（B） be the domain of the double sequential band matrix B（r, s） in the sequence spaces f0 and f. In this article, the β- and γ-duals of the space f（B） are determined. Additionally, we give some inclusion theorems concerning with the spaces f0（B） and f（β）. Moreover, the classes （f（B） ： μ） and （μ： f（B）） of infinite matrices are characterized, and the characterizations of some other classes are also given as an application of those main results, where μ is an arbitrary sequence space.
