In the field of machine learning, deep generative models have been one of the most interesting research directions. In this project, we studied an existing generative model that attempts to solve the problem of translating object photos into hand-drawn sketches. The model consists of convolutional neural networks for encoding and decoding pixel-based photos and recurrent neural networks for encoding and decoding vector-based sketches. We evaluated the output of the model. The result is that the model is good at translating pixel-based sketch images to vector-based sketch, but it has difficulties generating recognizable sketches from actual photos. We analyzed possible problems and discussed approaches to fix and improve the model in future work. The model has good potentials for learning pixel-based photos of objects and generating vector-based sketches stroke by stroke.