v2026.11.2
All Bundles
Bundle

Convolutional neural networks on the CPU: layers with forward and backward passes, the Adam and SGD optimizers, an image dataset with train, validation and test splits and on-the-fly augmentation, and a training loop that records the metrics of every epoch. The layers are native (libobjk_ml, Eigen plus a small thread pool); this library holds the model API, the dataset and the trainer. Compile with -lib ml_cnn,ml,ml_lab,opencv,csv,json,cipher,gen_collect. Design, validation and measurements: docs/AI.md, "Convolutional networks (System.ML.Cnn)".

Initializer

How the weights of a Conv and Dense layer start out. GLOROT is Glorot (Xavier) uniform, the default of Keras: limit sqrt(6 / (fan_in + fan_out)). KAIMING is He uniform, limit sqrt(6 / fan_in), for ReLU networks. Biases always start at zero. The draws come from the model's seed, so the same seed gives the same weights on every platform.

Items