determine malware or not using RandomForest and 3 layer deep learning model

using scikit-learn to determine malware or not using Random Forest
Random forests model is consists of 64 tree and input vector size is 20000 by hashing tricks
also use keras to create deep learning model ( input 1024 vector, 2 internal layer consists of 512 and 64 ReLU node and sigmoid output ) to check malware.