Version 2.1.1

• Fix some minor issues
• Adding more tests

Version 2.1.0

• More than 30% improvement in final results
• Using new model for Persian g2p-seq2seq prediction
• Detecting Persian Kaser Ezafe (Genitive detection)
• More compatible with Hazm tag-sets
• Using espeak-ng with Farsi language
• Updating Persian affixes
• Homograph disambiguation based on their POS and weight, [examples]
• Breaking unknown words
• Simplifying lexicon results
• Dockerfile is added
• Adding more tests cases
• Updating lexicon
• Fixing some bugs and issues

Version 2.0.0

• Add grpc_server tools to run tihu on a server
• Update functions and callback
• Load all synthesizer (voices) at the beginning
• Update tools (console and play)
• Fix travis errors
• fix some issues on looking up word
• Using tihu tokenizer to pars the text
• A word can have more than one entry
• Make corpus module
• reorganize the folders
• Fixing some bugs and issues