“FreeNeb Status Report 2016-09-05”版本间的差异

Coding done RNN-TTS model, and test it.

Learn HTS source code.

• wolf09 is ready.
• Finish backup of corpus0.
• Start 11000h-16k TDNN training. Training process is about 0.6 iter.
• Test vivi-decoder.

• Learn something about vivBeam.
• Data preparation and feature extraction for 8k speech.

• Finished nnet3 decoder on python server

• Do some investigation on CuDNN

[WD] vivEngine doc?