1 个仓库
Building complex covariance kernels by combining simpler kernels through mathematical operations.
Distinct from Kernel Composition Frameworks: Distinct from GPU kernel implementation by focusing on the statistical composition of covariance functions.
Explore 1 awesome GitHub repository matching artificial intelligence & ml · Kernel Composition. Refine with filters or upvote what's useful.
GPyTorch 是一个 GPU 加速的概率框架和 PyTorch 库,用于实现可扩展的高斯过程模型。它提供了一套用于高斯过程建模和不确定性估计的系统,旨在在图形硬件上执行高效的矩阵运算。 该框架具有一个用于构建自定义协方差函数和建模复杂数据依赖关系的模块化内核系统。它特别将高斯过程与深度神经网络集成,以创建用于回归和分类的混合模型。 该系统采用数值线性代数技术,包括预处理共轭梯度和基于张量的运算,以处理大型数据集。它还支持黑盒变分推理和用于超参数优化的自动微分。
Allows the construction of complex covariance structures by combining simple kernels through addition and multiplication.