Closed taloglu closed 10 months ago
My mistake
Change line 363 to model = WhisperModel(model_size, device="cpu")
instead of cuda and float16.
@taloglu have you tested how much time it takes? showing me 5 hours for 7 min video : https://github.com/fictions-ai/autocaption/issues/4
I get the following error when I run the Streamlit application.
This CTranslate2 package was not compiled with CUDA support
Would you know what I should do to fix it? Do I need a GPU to run the model? I am using a Macbook Pro M1.