phase 2 Neural networks

[MahdiEsrafili]

• [x] more data augmentation and transforms
• [x] make intel classification network bigger
• [x] plot loss after training
• [x] use tensorboard for monitoring
• [x] use learning rate scheduler
• [x] use batch normalization and drop out layer in network
• [x] sample data for training
• [x] user pytorch DatasetFolder to create Intel image dataset. link
• [x] user set_description for tqdm to print loss and other info

[MahdiEsrafili]

• [x] use sklearn classification_report to obtain precision and recall

[MahdiEsrafili]

• [x] for name classification with RNN, show data distribution over classes
• [x] show line2tensor a name (I mean, show how is the representation of a name as pytorch tensor)