1 个仓库
Sequence models that utilize stacks of convolutional layers instead of recurrent units for parallel processing.
Distinct from Sequence-to-Sequence Models: Distinct from general Sequence-to-Sequence Models: specifically focuses on the convolutional architecture rather than encoder-decoder patterns.
Explore 1 awesome GitHub repository matching artificial intelligence & ml · Convolutional Sequence Models. Refine with filters or upvote what's useful.
TCN 是一个深度学习序列框架和库,用于构建时间卷积网络。它提供了一个工具包,用于实现纯卷积架构来建模序列数据,作为循环神经网络 (RNN) 的替代方案。 该项目包括一个序列建模基准测试套件,旨在评估架构的准确性和处理速度。该套件利用标准化任务,包括记忆问题、数字分类、音乐和语言任务,来量化性能。 该框架涵盖了用于序列处理的一系列结构组件,例如防止信息泄露的因果卷积层和增加感受野的扩张卷积核。它还提供用于梯度稳定性的残差连接块,以及用于将原始时间序列数据转换为张量的数据管道。
Replaces recurrent neural networks with a stack of convolutional layers to allow parallel processing of sequences.