[Caligula]

Any chance of releasing the training data you used? Also what are the plans with DeepSpeech? Just for use by baidu or will it be released as open source or a developer api service?