v0.2.0

• Update documentation.
• Add more models for NLP tasks (albert-tiny, albert-base, bert-base)
• Improve tokenizer. [#2]
• Add arguments for GPU usage, batch usage.
• Add support to sentence segmentation.