Thesis Project This is for self preservation purpose, therefore expect no updates of codes. The goal is to build a detection model that can clarify hate speech from various Malaysian social media posts. GRU and LSTM performs the best, without any tokenization method.