ColdFir5/Uberduck-Tacotron2-Multi-Speaker-GSTs-Implementation

Currently this is not working as the repo used to make the notebook work is under-development, This is a notebook from Kaggle I had made that allows user's to make their own AI voices using 16bit PCM, 22050 HZ WAV files on the Neural networks provided by NVIDIA's creation of Tacotron 2 which has been further developed and worked on by the team at Uberduck.ai in order to add other amazing features to it using Tacotron 2 GPT. These AI TTS voices can be used regularly to output synthesised vocals or you can use reference audio in order to let the AI adapt to the flow of the reference audio to output a similar result.