Optimisers

Optimisers (or optimisers) are algorithms that adjust neural network weights to minimise the loss function during training. The athena library implements several popular optimisation algorithms, each with different characteristics and use cases.

Available Optimisers

• SGD Optimiser
• Adam Optimiser
• RMSprop Optimiser
• Adagrad Optimiser

Creating custom optimisers

The athena library is designed with extensibility in mind, allowing users to create custom optimisers by extending the base_optimiser_type.

See the tutorial: Creating Custom Optimisers