ML_palyground

• Bioresponse(RF and GB) - predicting biological response of a molecule based on tabular data from Kaggle competition. Used and compared: Random Forest and Gradient Boosting models. Additionally, I explored the effect of the size of train data on the quality of models via learning_curve.
• AE_MNIST - implementing Autoencoder and Variational Autoencoder architectures for reconstruction and generation handwritten digits from MNIST dataset.
• AE_face_reconstruction - implementing Autoencoder and Variational Autoencoder architectures for face reconstruction and face generation. Dataset: Labeled faces in the Wild
• GAN_Flickr-Faces - implementing GAN architecture for working with 512x512 faces from Flickr-Faces dataset.