How to run pre-trained models to generate voice e.g. Joe Rogan?

[zarifaziz]

Hi, Can you please provide instructions on how to use this repo and generate speech from input text? Sorry I'm just having a bit of trouble figuring out how to run it.

Also is there a pre-trained model of Joe Rogan? How would I go about building this? Thanks!

[NoahZu]

+1

[Traincraft101]

If I remember correctly, you extract the pretrained voice into the logdir folder and run "python synthesize.py" in the command prompt. You can edit the "harvard_sentences.txt" file to make it say whatever you want. (just don't delete the numbers) Then a folder called "samples" is generated in the dc_tts-master folder with the wav files.

As for how to make a voice, I used the LJ speech dataset as a template to create my own dataset. I couldn't train it because I don't have enough ram at the moment, so I'll have to wait to see if it works.

edit: If you just want to mess around with creating/cloning voices, I would recommend this model. You don't need to invest a ton of time into a single voice.