fishaudio

/

drama-3

Like 1.2k

Follow

Fish Audio 2.13k

Thanks so much for making Drama-3 openly available!

We’ve added experimental Drama-3 support to the CTranslate2 inference engine:
https://opennmt.net/CTranslate2/guides/drama3.html#drama-3

This lets you run Drama-3 efficiently on both CPUs and GPUs, including modest hardware. It’s an initial implementation, and some features (e.g. voice cloning) aren’t supported yet.

We’re also releasing faster-tts, a new library that makes it easy to use CTranslate2 for TTS:
https://github.com/jordimas/faster_tts

pip install faster-tts

from faster_tts import Model

model = Model()
result = model.synthesize(
    "Hello! Welcome to our application.",
    output="hello.wav",  # Save the audio as a WAV file.
)

Feedback is very welcome. Please report issues on the relevant tracker: CTranslate2 for engine or conversion problems, faster-tts for library or usage problems.

Happy hacking!
Jordi