Learning Deep Representations of Fine-grained Visual Descriptions - View it on GitHub