MartinuzziFrancesco/RecurrentLayers.jl

AI Summary: Implement a new recurrent neural network layer, specifically the Update Gate RNN (UGRNN), for the RecurrentLayers.jl package in Julia. This involves translating the provided mathematical equations into Julia code, ensuring compatibility with the existing Flux.jl framework, and adding thorough testing.

AI Summary: Implement more specific initializers for the recurrent neural network layers in RecurrentLayers.jl, mirroring the approach used in Lux.jl and LuxRecurrentLayers.jl to improve the flexibility and control over model initialization. This involves modifying existing layer implementations to accept and utilize more detailed initializer specifications.