Edit model card


Text-to-Speech (TTS) with Transformer trained on LJSpeech

This repository provides all the necessary tools for Text-to-Speech (TTS) with SpeechBrain using a Transformer pretrained on LJSpeech.

The pre-trained model takes in input a short text and produces a spectrogram in output. One can get the final waveform by applying a vocoder (e.g., HiFIGAN) on top of the generated spectrogram.

Perform Text-to-Speech (TTS) - Running Inference

To run model inference pull the interface directory as shown in the cell below

Note: Run on T4-GPU for faster inference

!pip install --upgrade --no-cache-dir gdown
!gdown 1oy8Y5zwkLel7diA63GNCD-6cfoBV4tq7
!unzip inference.zip
%%capture
!pip install speechbrain
%cd inference
import torchaudio
from TTSModel import TTSModel
from IPython.display import Audio
from speechbrain.inference.vocoders import HIFIGAN

texts = ["This is a sample text for synthesis."]

model_source_path = "/content/inference"
# Intialize TTS (Transformer) and Vocoder (HiFIGAN)
my_tts_model = TTSModel.from_hparams(source=model_source_path)
hifi_gan = HIFIGAN.from_hparams(source="speechbrain/tts-hifigan-ljspeech", savedir="tmpdir_vocoder")

# Running the TTS
mel_output = my_tts_model.encode_text(texts)

# Running Vocoder (spectrogram-to-waveform)
waveforms = hifi_gan.decode_batch(mel_output)

# Save the waverform
torchaudio.save('example_TTS.wav',waveforms.squeeze(1), 22050)
print("Saved the audio file!")

If you want to generate multiple sentences in one-shot, pass the sentences as items in a list.

Inference on GPU

To perform inference on the GPU, add run_opts={"device":"cuda"} when calling the from_hparams method.

Note: For Training the model please visit this TTS_Training_Inference notebook

Limitations

The SpeechBrain team does not provide any warranty on the performance achieved by this model when used on other datasets.

About SpeechBrain

Citing SpeechBrain

Please, cite SpeechBrain if you use it for your research or business.

@misc{speechbrain,
  title={{SpeechBrain}: A General-Purpose Speech Toolkit},
  author={Mirco Ravanelli and Titouan Parcollet and Peter Plantinga and Aku Rouhe and Samuele Cornell and Loren Lugosch and Cem Subakan and Nauman Dawalatabad and Abdelwahab Heba and Jianyuan Zhong and Ju-Chieh Chou and Sung-Lin Yeh and Szu-Wei Fu and Chien-Feng Liao and Elena Rastorgueva and François Grondin and William Aris and Hwidong Na and Yan Gao and Renato De Mori and Yoshua Bengio},
  year={2021},
  eprint={2106.04624},
  archivePrefix={arXiv},
  primaryClass={eess.AS},
  note={arXiv:2106.04624}
}
Downloads last month
3
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social visibility and check back later, or deploy to Inference Endpoints (dedicated) instead.