Disco Core

The 'core-entities' module offers foundational components for building machine learning models, including tensors, activation functions, layers, networks, cost functions, and optimizers. It supports various activation functions like ReLU, Logistic, Tanh, GELU, Swish, and SwiGLU, as well as layers such as Dense and Convolutional. The module also provides network architectures like Sequential and Residual Networks, along with cost functions like Mean Squared Error and Cross Entropy Loss. Optimizers include Gradient Descent, Stochastic Gradient Descent, and Adam. Training components like Sequential Trainer, AutoDiff Trainer, and Organic Trainer are also part of this module.