Natural Language Processing Model Parameter Selection
Advanced Modeling PracticeWhen performing research on NLP models encoders, embedding layers, Regularization, LSTM and RNN layers, Convolutional layers as well as dense layers, and number/ratio of neurons all impact model training convergence, accuracy, and fit (or overfitting), but manually designing, compiling, training, and evaluating model's can be time consuming. In this research paper we examine the uses of a programmatic solution to automate the process.