NVIDIA / NeMo

A scalable generative AI framework built for researchers and developers working on Large Language Models, Multimodal, and Speech AI (Automatic Speech Recognition and Text-to-Speech)

https://docs.nvidia.com/nemo-framework/user-guide/latest/overview.html

Apache License 2.0

12.05k stars 2.51k forks source link

In notebook colab "TTS Inference Prosody Control" an error has been made. #3938

Closed Nikuson123 closed 2 years ago

Nikuson123 commented 2 years ago

In notebook colab: https://colab.research.google.com/github/NVIDIA/NeMo/blob/stable/tutorials/tts/Inference_DurationPitchControl.ipynb

an error has been made.

instead of:

Load the models from NGC

fastpitch = FastPitchModel.from_pretrained("tts_en_fastpitch").eval().cuda() hifigan = HifiGanModel.from_pretrained("tts_hifigan").eval().cuda()

Should be:

Load the models from NGC

fastpitch = FastPitchModel.from_pretrained("tts_en_fastpitch") hifigan = HifiGanModel.from_pretrained(model_name="tts_hifigan")

redoctopus commented 2 years ago

The current notebook code with .eval().cuda() works when I try it on Colab. Are you sure you've connected to an instance with a GPU? (Runtime -> Change runtime type -> select "GPU" under "Hardware accelerator")

Nikuson123 commented 2 years ago

Текущий код блокнота .eval().cuda()работает, когда я пробую его в Colab. Вы уверены, что подключились к экземпляру с GPU? (Среда выполнения -> Изменить тип среды выполнения -> выберите «GPU» в разделе «Аппаратный ускоритель»)

I definitely connected with the GPU, the problem was in the model name