search

metavoiceio / metavoice-src

Foundational model for human-like, expressive TTS
https://themetavoice.xyz/
Apache License 2.0
3.77k stars 652 forks source link

Kaggle Notebook? #142

Open MethanJess opened 4 months ago

MethanJess commented 4 months ago

Kaggle offers a free 16gb VRam GPU / TPU for 20-30 hours per week

It would be really nice to have a Metavoice notebook for it, I tried to import the Google Colab notebook into Kaggle but it kept throwing errors... Can anyone make a notebook for it? thanks!

Sarwarkhann commented 4 months ago

Hi

chinmaya-growexxer commented 1 month ago

Cloning the repo

!git clone -b vatsal/make_2_5x_faster https://github.com/metavoiceio/metavoice-src.git %cd metavoice-src

Install dependencies

!pip install -r requirements.txt !pip install --upgrade torch torchaudio !pip install -e .

Loading Text-to-Speech model --Metavoice

from fam.llm.fast_inference import TTS tts = TTS() response = "This is the demo for Metavoice TTS with voice cloning" wav_file = tts.synthesise( text=response, spk_ref_path="/kaggle/input/sample.mp3" ) print(f"Response generated:/n{display(Audio(wav_file, autoplay=True))}")

The above code is working properly in Kaggle notebook