v0.0.2 - Public release tweaks

Some minor changes to docs, README, etc.

Initial public release

Initial public release includes the following models:

• BERT (training, prediction, embeddings)
• MT-DNN (training, prediction, embeddings)
• Universal Sentence Encoder (embeddings)
• fastText (training, prediction, embeddings)
• pytorch_transformer Transformers (training, prediction, embeddings)

Additional features include:

• Experiments (ClassificationExperiments only)
• Data augmentation (Word2Vec, WordNet, and BERTMaskedLM)
• Document windowing