Supertone

/

supertonic-2

like 304

Follow

Supertone 286

Has anyone tested this on windows machines and not an M4 pro? Can it run cpu only? If not, what's the minimum vram?
@jujutechnology

738.13 ms, start to end (this does not include loading model, which takes about a second)
using CPU

import datetime
from supertonic import TTS
import os
os.environ["CUDA_VISIBLE_DEVICES"] = ""
tts = TTS(auto_download=True)
style = tts.get_voice_style_from_path("supertonic/py/assets/voice_styles/Pixie.json") # obv replace with your voice style, this is a custom mixed one

text = "This morning, I took a walk in the park, and the sound of the birds and the breeze was so pleasant that I stopped for a long time just to listen."
start = datetime.datetime.now()
wav, duration = tts.synthesize(text, voice_style=style, speed=1.2)

end = datetime.datetime.now()
length = (start-end).microseconds
print(length/1000, "ms, start to end")
tts.save_audio(wav, "results/out.wav")