Masterthesis - Predict a permanent magnet synchronous motor's components' temperature with LSTM/GRU recurrent neural networks using Chainer and particle swarm optimization.
This project is a fork of the chainer project (thats why it has so many contributors). The framework was not altered though.
Hyperparameter optimization of the used recurrent neural networks was conducted by a self-implemented particle swarm optimization. Moreover, it was automated on a computing cluster my university was providing (PC²).