![]() The YouTuber Negan Jeffrey posts SpongeBob song cover videos on YouTube that he has made with uberduck's SpongeBob SquarePants (regular) TalkNet model. Speaking of AI's SpongeBob Tacotron 2 model is a private voice model that has been created exclusively for SpongeDubs by reason of the fact that SpongeDubs is the one who gathered the training data for the model however, this rule is not applicable to the dataset that has been used to train the model, which Speaking of AI freely gave to Gosmokeless28 on May 20, 2021. Speaking of AI's SpongeBob Tacotron 2 model was trained on an LJ Speech base model, and it is probable that the neural vocoder for the voice model is a WaveRNN vocoder that was trained on the LJ Speech dataset, considering the small amount of reverb artifacts that are present in the output audio generated with the model. The training of the model was done during the mid-to-late part of November 2019. ![]() Speaking of AI made changes to the voice clips-in terms of the way they were arranged-to ensure that the dataset would be compatible for the model, and the dataset's transcript was written by both an auto-transcriber and by Speaking of AI himself. zip files from The Sounds Resource that contained SpongeBob voice clips that had been sourced from several video games. ![]() In the process of the creation of the first ever SpongeBob AI voice model, SpongeDubs downloaded. On Nov 12, 2019, SpongeDubs announced that he was going to get in touch with Speaking of AI (who is a YouTuber and AI researcher) with the intention of collaborating with him to create a Tacotron 2 model of SpongeBob's voice for the purpose of using it to generate audio clips of SpongeBob reading aloud song lyrics, which would be used to aid SpongeDubs in the process of him making SpongeBob sing songs. ![]()
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |